231 PM Skills

Use when you're preparing a 1:1 and want it to be a real conversation, not a status readout — framed around the one outcome it must produce, with the 2–3 topics that need live discussion prepped. Not for prepping a general meeting (`meeting-prep`), a standalone difficult conversation (`difficult-conversation`), or managing up as an ongoing channel (`manage-the-upward-channel`).

Category

Meetings & Facilitation

Output

Use when you're preparing a 1:1 and want it to be a real conversation, not a status readout — framed around…

Run

/pm-os:1-on-1

Use when you want to be coached through building a talk yourself, step by step — outline, storyboard, headlines, fill, rehearse — with a story structure underneath. Not for generating the narrative for you (`presentation-narrative`), a slide-by-slide outline (`problem-deck`), or stress-testing a finished deck (`exec-deck`).

Category

Communication, Writing & Influence

Output

Use when you want to be coached through building a talk yourself, step by step — outline, storyboard,…

Run

/pm-os:5-step-story-deck

Use when someone with ADHD is stuck on a task and needs a motivation plan built from the Four Cs (Captivate, Create, Compete, Complete) — a situation in, a tactic per C and 2-3 immediate action steps out. Not for tracking motivation patterns over time (`motivation-journal`), or a founder's focus-and-execution plan (`founder-focus`).

Category

Coaching, Self-Management & Career

Output

Use when someone with ADHD is stuck on a task and needs a motivation plan built from the Four Cs (Captivate,…

Run

/pm-os:adhd-motivation

Use when a PM wants to rehearse a hard conversation with a designer or engineer before it happens, or replay one that already went badly — a scenario in, a 3-5 exchange live roleplay plus an in-character debrief and one usable reframe out. `/coaching` mode 3. Not for debriefing something already handled (`situation-retrospective`), prepping a specific planned hard conversation (`difficult-conversation`), or reacting to an incident happening right now (`respond-under-fire`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants to rehearse a hard conversation with a designer or engineer before it happens, or replay…

Run

/pm-os:adversarial-roleplay

Use when a design challenge needs concrete interaction ideas grounded in Don Norman's affordance/signifier vocabulary — a stated challenge in, affordances, perceived affordances, and signifiers out, each tied to a specific user interaction. Not for critiquing an existing design (`product-design-analyzer`), or correcting UX terms in written requirements (`ux-terminology`).

Category

Design & Prototyping

Output

Use when a design challenge needs concrete interaction ideas grounded in Don Norman's affordance/signifier…

Run

/pm-os:affordances-signifiers

Use when you have usage/retention data and an app description and need to find the specific action or condition that predicts long-term engagement — the moment users first get the product's value. Feeds `growth-strategy`'s loop diagnosis and `onboarding-redesign`'s transformation design. Not for building the growth strategy around it (`growth-strategy`), redesigning the onboarding path to reach it faster (`onboarding-redesign`), or diagnosing why users churn after activation (`churn-reduction`).

Category

Discovery & Research

Output

Use when you have usage/retention data and an app description and need to find the specific action or…

Run

/pm-os:aha-moment

Use when an AI feature needs a one-page go/no-go before anyone builds it — a customer problem and one persona in, a canvas out chaining business outcome → product outcome → solution hypothesis → success metric, plus the data and model feasibility check, the cheap discovery acts that would de-risk it, and risks split into investigate-now versus monitor-later. Not for turning a problem into a single falsifiable hypothesis (`product-hypothesis`), designing the experiment that tests one (`experiment-design`), or building the KPI arithmetic behind a feature's impact (`impact-model`).

Category

Discovery & Research

Output

Use when an AI feature needs a one-page go/no-go before anyone builds it — a customer problem and one persona…

Run

/pm-os:ai-recommendation-canvas

Use when you want to be tutored on a topic *Socratically* — guided to the answer with questions and hints tailored to your level and prior knowledge, not handed the solution. Not for a structured mastery plan for a skill (`skill-mastery`), or PM performance coaching (`coaching`).

Category

Coaching, Self-Management & Career

Output

Use when you want to be tutored on a topic *Socratically* — guided to the answer with questions and hints…

Run

/pm-os:ai-tutor

Use when a new app needs its design *foundation* set before any screens get built — the one core feature, a grayscale sketch of the main screen, and a deliberately tiny design system. Fires at project kickoff or "where do I start designing this?". Not for full wireframe flows (`wireframe-sketch`), a testable prototype (`clickable-prototype`), or translating a PM spec to design (`pm-to-design`).

Category

Design & Prototyping

Output

Use when a new app needs its design *foundation* set before any screens get built — the one core feature, a…

Run

/pm-os:app-design-foundation

Surface, prioritize, and find early signals for the riskiest assumptions in your product strategy — from assumption generation through prioritization to identifying the easiest validation signal. Use when mapping product assumptions, prioritizing risk, or designing validation experiments.

Category

Discovery & Research

Output

Surface, prioritize, and find early signals for the riskiest assumptions in your product strategy — from…

Run

/pm-os:assumptions

Use when a PM wants a comprehensive read on the gaps they can't see from the inside — answers to a 12-question behavioral interview across all 7 excellence clusters in, 2-4 ranked blind spots out, each with why it's invisible, what it costs, and one observable change. `/coaching` mode 5. Not for debriefing one situation (`situation-retrospective`), one decision's process (`decision-audit`), or how the team specifically perceives them (`team-perspective-reveal`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants a comprehensive read on the gaps they can't see from the inside — answers to a…

Run

/pm-os:blind-spot-scan

Use when the user wants an ideation session on any topic where the point is to beat the *banal bar* — the obvious ideas anyone would produce — through breadth techniques: association, inversion, adjacent-field transplants. Fires on "brainstorm X with me," "I need non-obvious ideas about Y." Not for product-idea sessions (`product-brainstorm`), constraint-driven ideation (`constrained-ideas`), question generation (`good-question-brainstormer`), or problem reframing (`lateral-thinking`).

Category

Ideation & Analytical Thinking

Output

Use when the user wants an ideation session on any topic where the point is to beat the *banal bar* — the…

Run

/pm-os:brainstorm-genius

Reflects on breadboards by comparing expected behavior against implementation wiring and finding design smells. Use when reviewing or improving breadboards created by the breadboarding skill.

Category

PRD, Specs & Documentation

Output

Reflects on breadboards by comparing expected behavior against implementation wiring and finding design smells

Run

/pm-os:breadboard-reflection

Transform a workflow description into affordance tables showing UI and Code affordances with their wiring. Use to map existing systems or design new ones from shaped parts.

Category

PRD, Specs & Documentation

Output

Transform a workflow description into affordance tables showing UI and Code affordances with their wiring

Run

/pm-os:breadboarding

Use when a stakeholder hands over a vague *brief* — "make it modern," "improve the UX," "simplify this" — and the PM needs a concrete problem statement with a metric before any design work starts. Not for interpreting exec feedback on existing work (`exec-feedback`), framing from a session transcript (`problem-framing-canvas`), or pulling a team out of solution mode (`problem-first`).

Category

Decisions & Prioritization

Output

Use when a stakeholder hands over a vague *brief* — "make it modern," "improve the UX," "simplify this" — and…

Run

/pm-os:brief-to-problem

Use when a reported bug needs a fast now-or-later call against what's already committed — a bug description, current work, and planned work in, a defer-or-escalate decision out with reasoning against both. Not for classifying a decision's reversibility (`two-way-door`), or pricing a mid-project scope-creep request (`scope-defense`).

Category

Decisions & Prioritization

Output

Use when a reported bug needs a fast now-or-later call against what's already committed — a bug description,…

Run

/pm-os:bug-triage

Use when someone shares career thoughts and wants tailored guidance — themes reflected back from their own words, a path direction with trade-offs, positioning help, and one concrete *next step*. Not for writing the résumé itself (`resume`), interview preparation (`pm-interview`), building interview stories (`star-stories`), or a plan to master a specific skill (`skill-mastery`).

Category

Coaching, Self-Management & Career

Output

Use when someone shares career thoughts and wants tailored guidance — themes reflected back from their own…

Run

/pm-os:career-guidance

Use when a decision needs its root cause traced backward and each candidate option's downstream consequences traced forward, one branch at a time — recursive "why" until you hit bedrock, then forward from the live options to what each one actually causes. Also /decisions Step 1, producing the causal map and consequence tree the rest of the workflow builds on. Not for checking whether a list of options is MECE (`mece-tree`), the full Why-tree pipeline with testable hypotheses (`mckinsey-issue-tree`), or converging a messy problem to one recommendation (`structure-problem`).

Category

Ideation & Analytical Thinking

Output

Use when a decision needs its root cause traced backward and each candidate option's downstream consequences…

Run

/pm-os:causal-tree

Use when churned-user data and exit-survey responses need turning into a targeted reduction plan — each strategy tied to a specific stated reason for leaving, with a timeline and success metric. Not for the broader growth-loop diagnosis (`growth-strategy`), finding the activation moment that predicts retention (`aha-moment`), or reducing flow friction generally (`friction-reduce`).

Category

Metrics & Measurement

Output

Use when churned-user data and exit-survey responses need turning into a targeted reduction plan — each…

Run

/pm-os:churn-reduction

Use when a meeting has multiple attendees who each need a different influence approach — a hidden-agenda map (or a single target and goal) in, a Cialdini-principle-based strategy out, tailored per attendee to their specific motivation. Also `/meeting` Step 2, taking Step 1's hidden agenda map as input. Not for surfacing what attendees actually want first (`hidden-agendas`), analyzing your own delivery afterward (`leadership-presence`), or restructuring what you say for clarity (`pyramid-principle`).

Category

Communication, Writing & Influence

Output

Use when a meeting has multiple attendees who each need a different influence approach — a hidden-agenda map…

Run

/pm-os:cialdini

Use when someone calls a thing immeasurable — quality, engagement, security, brand, innovation — and it has to become countable anyway: three *links* from detectable consequence to observable amount to a tracking method. Also /measure Step 1, feeding the observables the rest of the workflow decomposes. Not for breaking an already-named quantity into estimable factors (`fermi-decomposition`), pricing whether the measurement is worth running (`value-of-information`), or setting one primary metric for a specific change (`success-metric`).

Category

Metrics & Measurement

Output

Use when someone calls a thing immeasurable — quality, engagement, security, brand, innovation — and it has…

Run

/pm-os:clarification-chain

Use when you have a screen sketch or image description and need a working, clickable HTML/CSS/JS prototype for user testing — built incrementally so each layer can be checked before the next. Takes a `wireframe-sketch` or `app-design-foundation` sketch as input. Not for the sketch itself (`wireframe-sketch`), or converting an existing screenshot into structured data (`ui-to-json`).

Category

Design & Prototyping

Output

Use when you have a screen sketch or image description and need a working, clickable HTML/CSS/JS prototype…

Run

/pm-os:clickable-prototype

Use when a change has to be announced and neither the framing nor the rollout is settled — two or three viable frames, an IC draft and an exec draft in the user's own voice, then a *sequenced* plan naming channels, owners, what to leave out, and how you'll know it landed. Also /stakeholder Step 4, taking the stakeholder map and risk register as input. Not for a severe outage needing an incident cadence and post-mortem (`crisis-comms`), a recurring progress update (`status-update`), or critiquing a draft that is already written (`exec-update-review`).

Category

Communication, Writing & Influence

Output

Use when a change has to be announced and neither the framing nor the rollout is settled — two or three…

Run

/pm-os:comms-plan

Use when ideas must fit hard *constraints* — a fixed budget, a project type, a deadline, a one-person team — and the constraints should drive the ideation rather than filter it afterward. Fires on "what can I do with $500," "ideas for X within Y." Not for open-topic ideation (`brainstorm-genius`), product-idea sessions (`product-brainstorm`), or reframing the problem itself (`lateral-thinking`).

Category

Ideation & Analytical Thinking

Output

Use when ideas must fit hard *constraints* — a fixed budget, a project type, a deadline, a one-person team —…

Run

/pm-os:constrained-ideas

Build a grounded product strategy from first principles — from identifying the strategic crux through competitive analysis to a limit-based strategy and value chain map. Use when developing product strategy, identifying the strategic crux, or doing competitive analysis.

Category

Strategy & Vision

Output

Build a grounded product strategy from first principles — from identifying the strategic crux through…

Run

/pm-os:strategy

Use when teams are stuck, decisions aren't being made, meetings are dysfunctional, or there's confusion about who has authority — a misalignment situation in (description, transcript, decision log, or vague complaint), a root-cause diagnosis and level-specific fixes out. Not for assigning decision rights going forward (`davci`), proposing a convergence point when there's no process or authority at all (`focal-point-finder`), or mapping per-person motivation instead of structural process (`hidden-agendas`).

Category

Stakeholders & Politics

Output

Use when teams are stuck, decisions aren't being made, meetings are dysfunctional, or there's confusion about…

Run

/pm-os:corporate-misalignment-finder

Use when you need a structured empathy map — thinks, feels, says, does, sees, hears, pains, goals — for one audience in relation to one topic, surfacing motivations and fears the audience wouldn't state outright. Feeds `persona-lens-decision`'s reaction analysis. Not for the end-to-end experience across stages (`journey-map`), an actionable task-level persona (`user-profile`), or building the first persona from research (`proto-persona`).

Category

Discovery & Research

Output

Use when you need a structured empathy map — thinks, feels, says, does, sees, hears, pains, goals — for one…

Run

/pm-os:create-empathy-maps

Use when a severe product outage needs a communication plan — stakeholders ranked by impact, a tailored message per group, a promised update *cadence*, and a post-mortem framework to fill after. Not for a routine communication plan (`comms-plan`), a launch announcement (`press-release`), or mapping stakeholder power and interest (`stakeholder-map`).

Category

Stakeholders & Politics

Output

Use when a severe product outage needs a communication plan — stakeholders ranked by impact, a tailored…

Run

/pm-os:crisis-comms

Use when expertise has to be extracted rather than asked for — Gary Klein's Critical Decision Method runs a 4-pass cognitive interview over one non-routine incident, surfacing the perceptual *cues* and hidden assumptions an expert cannot state when asked directly, and ends in decision requirements, a critical cue list, a debrief, and a training scenario. Runs on someone else or on your own past decision. Not for auditing one decision's process quality (`decision-audit`), debriefing a situation you just came out of (`situation-retrospective`), or coding a customer interview for JTBD forces (`interview-insights`).

Category

Discovery & Research

Output

Use when expertise has to be extracted rather than asked for — Gary Klein's Critical Decision Method runs a…

Run

/pm-os:critical-decision-method

Use when a product page needs call-to-action copy — a product in, 12 tight CTAs out across four proven strategies (match the feeling, actionable next step, handle the objection, make it specific), each 6-35 characters. Not for the underlying value proposition the CTA sits on top of (`positioning`), or restructuring existing text for logical clarity (`pyramid-principle`).

Category

Communication, Writing & Influence

Output

Use when a product page needs call-to-action copy — a product in, 12 tight CTAs out across four proven…

Run

/pm-os:cta-copy

Use when aiming to establish structured decision-making protocols with defined roles and responsibilities — a decision and its stakeholders in, a Decider/Approver/Veto/Consulted/Informed assignment out per decision object. Also `/decisions` Step 6, taking the Step 5 recommendation as the decision to assign rights for. Not for diagnosing why authority is unclear in the first place (`corporate-misalignment-finder`), recording the bet once rights are assigned (`decision-journal`), or converging without any explicit authority at all (`focal-point-finder`).

Category

Stakeholders & Politics

Output

Use when aiming to establish structured decision-making protocols with defined roles and responsibilities — a…

Run

/pm-os:davci

Use when a claim needs its rhetorical cover stripped to see whether anything is under it — framework name-drops ("classic cold-start problem"), coined terms ("engagement debt"), contrarian inversions, aphorisms as trump cards, altitude statements ("own the workflow, not the feature"), meta-escalations ("we're solving the wrong problem"), or speculative second-order chains — restated *plainly* and tested against what it would actually change. Fires on "de-clever this", "plain version of this", "is this clever or true?", and on strategy docs, PRDs, and decision rationale that lean on the moves above. Not for packaging an idea that already survived (`prep-the-room`), an exec update or date commitment (`manage-the-upward-channel`), or restructuring prose for readability (`skimmable-writing`). Source: Shreyas Doshi's "sound less clever" post.

Category

Communication, Writing & Influence

Output

Use when a claim needs its rhetorical cover stripped to see whether anything is under it — framework…

Run

/pm-os:de-clever

Use when a PM wants one decision they already made audited on process rather than outcome — the decision, its alternative and how it was communicated in, a read on its rigor, transparency and customer-centricity out, plus the one insight that matters most. `/coaching` mode 4. Not for logging a decision as it's being made (`decision-journal`), classifying reversibility before committing (`two-way-door`), or a systematic audit across all behaviors (`blind-spot-scan`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants one decision they already made audited on process rather than outcome — the decision, its…

Run

/pm-os:decision-audit

Use when a decision is about to be made and needs to go *on the record* — expected outcomes with probabilities, assumptions with base rates, quit conditions, and a review date — so the future review judges the process, not the hindsight. Also /decisions Step 3, taking the causal map and classification as input. Not for auditing a past decision (`decision-audit`) or classifying reversibility (`two-way-door`).

Category

Decisions & Prioritization

Output

Use when a decision is about to be made and needs to go *on the record* — expected outcomes with…

Run

/pm-os:decision-journal

Use when a PM needs a *story-driven* demo script for a prototype or feature — a named protagonist, a scenario with stakes, and a beat-by-beat walkthrough — instead of a feature tour. Fires on "script my demo," "make this demo land," or prep for showing a prototype to stakeholders. Not for the deck around the demo (`presentation-narrative`, `5-step-story-deck`) or for reading the room politically (`prep-the-room`).

Category

Design & Prototyping

Output

Use when a PM needs a *story-driven* demo script for a prototype or feature — a named protagonist, a scenario…

Run

/pm-os:demo-narrative

Use when one buyer's *switch* is stuck and needs diagnosing — a transcript, thread, or pipeline note in, their timeline stage, the four forces scored against evidence, a qualify/disqualify verdict, the dominant blocker, and the verbatim next move out. Also fires on prepping or debriefing a sales call. Not for coding one customer interview into JTBD forces (`interview-insights`), clustering forces across many interviews (`jtbd-forces`), or writing the job statements for a whole market (`jtbd-jobs`).

Category

Discovery & Research

Output

Use when one buyer's *switch* is stuck and needs diagnosing — a transcript, thread, or pipeline note in,…

Run

/pm-os:diagnose-the-switch

Use when you need to produce a one-page prep doc, provide calm opening lines and questions, anticipate reactions with de-escalation responses, and close with clear next steps for a planned hard conversation. Also `/stakeholder` Step 6, following Step 5's `meeting-prep` brief. Not for a reactive incident that just happened (`respond-under-fire`), a fast field-read before any interaction (`prep-the-room`), or scripting a public outage update (`crisis-comms`).

Category

Stakeholders & Politics

Output

Use when you need to produce a one-page prep doc, provide calm opening lines and questions, anticipate…

Run

/pm-os:difficult-conversation

Use when an incumbent company or product needs counter-positioning ideation — assumption-breaking "what if" questions spanning technology, business model, customer experience, and market that expose where a new entrant could win. Not for the table-stakes match/differentiate/leapfrog call (`parity-vs-differentiation`), portfolio-gap growth ideas (`pioneer-migrator-settler`), or the four-view adaptive plan (`scenario-plan`).

Category

Ideation & Analytical Thinking

Output

Use when an incumbent company or product needs counter-positioning ideation — assumption-breaking "what if"…

Run

/pm-os:disruption-what-ifs

Use when one specific edge case (typically from support or QA) needs a fix that doesn't overcomplicate the core product for everyone else — a scenario in, a recommended solution out, scored against the complexity tax it adds to the main experience. Not for sweeping a brief for the full set of edge cases upfront (`ux-edge-cases`), or turning the fix into testable QA acceptance criteria (`ui-acceptance-criteria`).

Category

PRD, Specs & Documentation

Output

Use when one specific edge case (typically from support or QA) needs a fix that doesn't overcomplicate the…

Run

/pm-os:edge-case-balance

Use when a PM wants application *archetypes* instead of a single answer or an internal-architecture call. Three openings: "I want to build X — what's out there?"; "I want to do X — what kind of app is this even?"; "they said it should be a Slack app — is it?". Returns 3 archetypes backed by real GitHub repos and a paragraph the user can paste to an AI builder. Not for an engineering team's internal architecture call (`tech-arch-brief`), bug triage (`bug-triage`), or translating a decided plan to stakeholders (`tech-to-business`).

Category

PRD, Specs & Documentation

Output

Use when a PM wants application *archetypes* instead of a single answer or an internal-architecture call

Run

/pm-os:eng-shape

Use when a build needs engineering in function-space rather than problem-space — symptoms reframed as measurable functions, the chain designed right-to-left, prototypes laid out as orthogonal arrays instead of A/B tests, scope cut to a kick-ass half, and a wall set. Not for shaping an idea into a pitch with an appetite (`shape-up`), finding the application archetype before building (`eng-shape`), holding scope once it is already under pressure (`scope-defense`), or a retrospective on a finished project alone (`situation-retrospective`).

Category

PRD, Specs & Documentation

Output

Use when a build needs engineering in function-space rather than problem-space — symptoms reframed as…

Run

/pm-os:engineer-for-progress

Use when you have a deck or investment case and need to stress-test it through the executive's skeptical lens before you present — surfacing the assumptions that must hold, the ways it blows up, the analyses that would settle it, and the questions to expect. Not for building the deck (`problem-deck`, `presentation-narrative`), planning it for an audience (`strategic-deck`), or writing a status update (`status-update`).

Category

Communication, Writing & Influence

Output

Use when you have a deck or investment case and need to stress-test it through the executive's skeptical lens…

Run

/pm-os:exec-deck

Use when an executive drops context-light feedback on existing work — "this feels cluttered," "not strategic enough," "something's off" — and the PM must act without a follow-up meeting on the calendar. Produces competing *interpretations*, a send-ready clarifying message, and a no-regret interim step. Not for a vague incoming brief (`brief-to-problem`) or preparing for the exec meeting itself (`prep-the-room`).

Category

Decisions & Prioritization

Output

Use when an executive drops context-light feedback on existing work — "this feels cluttered," "not strategic…

Run

/pm-os:exec-feedback

Use when an update or deck outline is drafted and needs a senior exec's read *before* you deliver it — a three-bullet TL;DR rewrite, an issue→fix pass over the vagueness and hedging, a clean narrative, and three-minute speaking notes. Also /stakeholder Step 7. Not for writing the update from scratch (`status-update`), planning the multi-channel rollout around it (`comms-plan`), or stress-testing an investment case against the objections it will meet (`exec-deck`).

Category

Communication, Writing & Influence

Output

Use when an update or deck outline is drafted and needs a senior exec's read *before* you deliver it — a…

Run

/pm-os:exec-update-review

Use when you have a hypothesis (or a goal to validate) and need the minimum experiment that could *disprove* it — method, primary metric, minimum detectable effect, sample size, and stop/scale rules. Also /research Step 5, one experiment per product hypothesis. Not for defining the success metric itself (`success-metric`), instrumenting the events an experiment needs (`tracking-schema`), or choosing which assumptions to test (`risky-assumptions`).

Category

Discovery & Research

Output

Use when you have a hypothesis (or a goal to validate) and need the minimum experiment that could *disprove*…

Run

/pm-os:experiment-design

Use when a shipped feature's results writeup is drafted and needs review before it circulates — checking whether the effect is real, whether confounders were ruled out, whether segment results are cherry-picked, and whether the recommendation actually *follows* from the numbers. Not for designing the experiment before it runs (`experiment-design`), diagnosing an unexplained fall in a live metric (`metric-drop`), or estimating a feature's impact before it's built (`impact-sizing`).

Category

Metrics & Measurement

Output

Use when a shipped feature's results writeup is drafted and needs review before it circulates — checking…

Run

/pm-os:feature-writeup

Use when a PM wants a post-launch learning plan — *checkpoints* after launch, what each one measures, who reviews it, and how findings reach the next iteration's plan. Also fires on "we launched X and never learned anything from it." Not for defining the success metric itself (`success-metric`), designing an experiment (`experiment-design`), or NPS programs (`nps-to-cx-plan`).

Category

Decisions & Prioritization

Output

Use when a PM wants a post-launch learning plan — *checkpoints* after launch, what each one measures, who…

Run

/pm-os:feedback-loop

Use when a number is needed and no data exists to look it up — market size, annual cost of a problem, how many users a change reaches — broken into factors that multiply or add, each carrying a 90% range so the *uncertainty* survives to the answer. Also /measure Step 2, decomposing the observables the clarification chain produced. Not for turning a vague intangible into observables first (`clarification-chain`), bounding a median from a handful of real samples (`rule-of-five`), or a funnel-based estimate of one feature's impact (`impact-sizing`).

Category

Metrics & Measurement

Output

Use when a number is needed and no data exists to look it up — market size, annual cost of a problem, how…

Run

/pm-os:fermi-decomposition

Use when a strategy problem is messy or underspecified and the PM needs the one *crux* — the structural reason the gap exists — found before any option is generated, then a recommendation stress-tested against it. Also `/strategy` Step 0, where it stops at the crux and hands off. Not for decomposing a problem into a full Why/What/How issue tree (`mckinsey-issue-tree`), writing the three-part strategy kernel (`strategy-kernel`), or MECE-checking a list you already hold (`mece-tree`).

Category

Strategy & Vision

Output

Use when a strategy problem is messy or underspecified and the PM needs the one *crux* — the structural…

Run

/pm-os:find-the-strategic-crux

Use when someone asserts a fixed constraint — "we can't", "it costs Z", "that's impossible", "it has to work this way" — or when an organizational problem is stuck with no obvious solution. Reduces the claim's key words to observable mechanisms, matches the problem to a known structure (public goods, principal-agent, queues, power laws), separates physics and economics from inherited convention, and returns priced options with a verdict. Researches concept anatomy, prices, counterexamples and base rates with whatever search tools are available. Also runs as a two-minute check on everyday claims: a standup estimate, a vendor quote, a compliance refusal. Use `causal-tree` instead when something has already happened and needs explaining, or `limit-strategy` when the limit state is the deliverable rather than a test. Not for pulling a team off a premature solution (`problem-first`), sizing a number on its own (`fermi-decomposition`), or improving by removing scope (`via-negativa`).

Category

Ideation & Analytical Thinking

Output

Use when someone asserts a fixed constraint — "we can't", "it costs Z", "that's impossible", "it has to work…

Run

/pm-os:first-principles

Use when a company's successes and disappointments need distilling into a Jim Collins-style *flywheel* — 4–6 components in a self-reinforcing loop that explains why the wins won and the losses lost. Fires on "what's our flywheel" or "why do some bets work for us and others don't." Not for the growth model and its constraint (`growth-strategy`), sequencing items into a loop (`reinforcing-sequence`), or durable moats (`moats`).

Category

Strategy & Vision

Output

Use when a company's successes and disappointments need distilling into a Jim Collins-style *flywheel* — 4–6…

Run

/pm-os:flywheel

Use when multiple parties need to converge on a single choice without full communication, such as setting a standard, deadline, or meeting point — a coordination problem in, a stress-tested focal point recommendation out. Not for assigning explicit decision rights when authority exists (`davci`), diagnosing why a decision process is stuck (`corporate-misalignment-finder`), or mapping stakeholders' power and interest (`stakeholder-map`).

Category

Stakeholders & Politics

Output

Use when multiple parties need to converge on a single choice without full communication, such as setting a…

Run

/pm-os:focal-point-finder

Use when a startup founder is scattered across too many priorities and needs to narrow to what matters, turn it into one simple trackable solution, and push the highest-intensity version of executing it. Not for ADHD-specific task motivation (`adhd-motivation`), or tracking your own energy patterns over time (`motivation-journal`).

Category

Coaching, Self-Management & Career

Output

Use when a startup founder is scattered across too many priorities and needs to narrow to what matters, turn…

Run

/pm-os:founder-focus

Use when a multi-step flow (onboarding, checkout, signup, approval) is losing people and you want to cut steps without removing the checkpoints that prevent costly errors — each step tagged, priced against the drop-off it causes, and redesigned. Not for the first-run activation path (`onboarding-redesign`), mapping the whole end-to-end experience (`journey-map`), or retention work (`churn-reduction`).

Category

Design & Prototyping

Output

Use when a multi-step flow (onboarding, checkout, signup, approval) is losing people and you want to cut…

Run

/pm-os:friction-reduce

Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a PPT/PPTX to web, or create slides for a talk/pitch. Helps non-designers discover their aesthetic through visual exploration rather than abstract choices.

Category

Communication, Writing & Influence

Output

Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files

Run

/pm-os:frontend-slides

Use when you have a use-case list (or a persona, goals, and product context, if no use cases exist yet) and need INVEST-compliant user stories with embedded Given/When/Then acceptance criteria — atomic, traceable, and covering the edge case as well as the happy path. Also `/prd` Step 4b — the Gherkin mode, alternative to Step 4a's plain `user-stories` — taking Step 3's use-case list as input. Not for plain role-goal-benefit stories with no embedded scenarios (`user-stories`), translating stories into testable UI QA specs (`ui-acceptance-criteria`), or defining the use cases themselves (`use-cases`).

Category

PRD, Specs & Documentation

Output

Use when you have a use-case list (or a persona, goals, and product context, if no use cases exist yet) and…

Run

/pm-os:gherkin-stories

Use when the user's own words show a bad-PM anti-pattern — a solution named before any problem, authority as the whole justification ("the CEO asked for it"), an excuse or a blamed team, "that's not my job", skipped validation, boilerplate or copied-competitor strategy, a roadmap of quick wins, a metric standing in for a judgment call, or a request to make a weak idea sound strategic — and one question should make them see it themselves. Also when they ask "am I being a bad PM?", and when /review or /daily-drip needs the good-PM catalog. Not for a full coaching session on a recurring pattern (`coaching`), auditing one decision's process (`decision-audit`), or stripping clever framing out of a claim (`de-clever`).

Category

Coaching, Self-Management & Career

Output

Use when the user's own words show a bad-PM anti-pattern — a solution named before any problem, authority as…

Run

/pm-os:good-pm-bad-pm

Use when a topic, decision or vague curiosity needs better questions before it gets answers — 25+ candidates generated across seven lanes (why, what if, how, eigenquestion, contextual, appreciative, story/life), then sharpened and ranked into a top 5-10 with a best-first question and a next action. Also `mckinsey-issue-tree` Phase 6, applied to the crux. Not for generating ideas rather than questions (`brainstorm-genius`), deliberately strange provocation questions (`offbeat-questions`), framing the problem statement itself (`problem-framing-canvas`), or resolving one ambiguous request through follow-ups (`clarification-chain`).

Category

Ideation & Analytical Thinking

Output

Use when a topic, decision or vague curiosity needs better questions before it gets answers — 25+ candidates…

Run

/pm-os:good-question-brainstormer

Use when a product's growth needs a strategy built from its actual growth model — how it acquires and keeps users today, the *constraint* currently limiting that loop, where the model runs out, and the methods that attack the constraint. Fires on "how do we grow this" or "growth has stalled." Not for building the flywheel diagram (`flywheel`), reducing churn specifically (`churn-reduction`), or finding the activation moment (`aha-moment`).

Category

Strategy & Vision

Output

Use when a product's growth needs a strategy built from its actual growth model — how it acquires and keeps…

Run

/pm-os:growth-strategy

Use when someone faces a personal opportunity or commitment and wants to check whether external pressure is overriding their *gut* — capture the instinct, name the pressures, weigh the honest costs, and see if the gut holds. Not for classifying a decision's reversibility (`two-way-door`) or the full product-decision workflow (`decisions`).

Category

Decisions & Prioritization

Output

Use when someone faces a personal opportunity or commitment and wants to check whether external pressure is…

Run

/pm-os:gut-check

Use when the user wants a live system tour, asks what PM OS can do, or says '/help' or '/pm-help'. Not for orienting on current state and a recommended next action (`pm-status`), matching one task to one skill (`pm-os-skill`), or browsing everything visually (`skill-browser`).

Category

PM OS System & Operations

Output

Use when the user wants a live system tour, asks what PM OS can do, or says '/help' or '/pm-help'

Run

/pm-os:pm-help

Use when multiple stakeholders' stated positions may not match what they actually want — a situation and stakeholder list in, a hidden-agenda map out (per-stakeholder likely motivation, fear, and stake). Also `/meeting` Step 1, feeding Step 2's `cialdini` influence design. Not for a fast field-read when short on time (`prep-the-room`), turning the map into influence tactics (`cialdini`), mapping formal authority instead of motivation (`power-map`), or reviewing your own delivery afterward (`leadership-presence`).

Category

Stakeholders & Politics

Output

Use when multiple stakeholders' stated positions may not match what they actually want — a situation and…

Run

/pm-os:hidden-agendas

Use when a PM wants to interrogate a product decision by talking *to* the ideal customer rather than about them — a first-person roleplay that answers only from specific recent episodes and refuses hypotheticals, aggregates, and speaking for other users. Not for a third-person analytical read of how one segment reacts to a decision (`persona-lens-decision`), synthesizing real interview transcripts (`interview-insights`), or generating a simulated JTBD interview transcript (`jtbd-transcript-sim`).

Category

Discovery & Research

Output

Use when a PM wants to interrogate a product decision by talking *to* the ideal customer rather than about…

Run

/pm-os:icp-roleplay

Use when a specific KPI number has to survive scrutiny — build the explicit arithmetic model of a feature's impact on named KPIs, every term a sourced or flagged assumption, with sensitivity ranges on the shaky ones. Not for a first rough is-it-worth-building estimate (`impact-sizing`), defining the success metric (`success-metric`), or designing the test that validates the assumptions (`experiment-design`).

Category

Metrics & Measurement

Output

Use when a specific KPI number has to survive scrutiny — build the explicit arithmetic model of a feature's…

Run

/pm-os:impact-model

Use when you need a top-down estimate of whether a feature is worth building — a funnel from exposed users down to actual adopters, converted into rough engagement and revenue impact, with the riskiest assumptions named. Not for a per-KPI arithmetic model that defends a number to execs (`impact-model`), defining the success metric for a change (`success-metric`), or the full measurement workflow (`/measure`).

Category

Metrics & Measurement

Output

Use when you need a top-down estimate of whether a feature is worth building — a funnel from exposed users…

Run

/pm-os:impact-sizing

Use when the user wants to import context from another AI assistant (Claude, ChatGPT, Gemini) via paste-back, has a saved memory dump to paste in, or says '/import-ai-memory'. Not for onboarding from scratch with guided questions (`pm-os-start`), explicit one-off capture of a decision or update (`pm-os-capture-memory`), or the slow one-question-a-day accretion loop (`pm-os-daily-drip`).

Category

PM OS System & Operations

Output

Use when the user wants to import context from another AI assistant (Claude, ChatGPT, Gemini) via paste-back,…

Run

/pm-os:import-ai-memory

Use when a company's strengths need to be pointed at a specific industry trend to find strategic options — not the full 7-part product strategy, just the trend/strength intersection. Not for the full structured strategy document (`product-strategy`), diagnosing the strategic problem itself (`strategy-kernel`), or the complete strategy workflow (`/strategy`).

Category

Strategy & Vision

Output

Use when a company's strengths need to be pointed at a specific industry trend to find strategic options —…

Run

/pm-os:industry-strategy

Use when forecasting future product opportunities from external shifts — a timeframe in, regulatory/technological/belief inflection points identified, their intersections found, and opportunities derived from those intersections out. Not for backcasting from an idealized end state of a specific product (`limit-strategy`), mapping where an existing product generates value (`value-chain`), or a general competitive scan (`netmba-competitor-analysis`).

Category

Strategy & Vision

Output

Use when forecasting future product opportunities from external shifts — a timeframe in,…

Run

/pm-os:inflection-scan

Use when one interface image needs describing in four different registers for four different audiences — novice instructions, a technical process flow, an ELI5 simplification, and a casual/social tone. Not for translating a PM's context into design requirements (`pm-to-design`), or reacting to a decision from one specific segment's lens (`persona-lens-decision`).

Category

Design & Prototyping

Output

Use when one interface image needs describing in four different registers for four different audiences —…

Run

/pm-os:interface-4-lenses

Use when one customer interview transcript needs JTBD analysis — the *four forces* (pushes, pulls, habits, anxieties) coded with verbatim quotes, plus the jobs, struggles, and workarounds tagged by type and intensity. Also /research Step 2. Not for coding many interviews into a clusterable dataset (`jtbd-forces`), diagnosing a stuck deal with the forces (`diagnose-the-switch`), or reconstructing a work conversation (`transcript-insights`).

Category

Discovery & Research

Output

Use when one customer interview transcript needs JTBD analysis — the *four forces* (pushes, pulls, habits,…

Run

/pm-os:interview-insights

Use when a cleaned interview transcript needs atomized, citable notes under a chosen framework (Chronological, Topical, AEIOU, or Empathy Map) — a transcript and a mode in, one-idea-per-note tables out, each note quoted and timestamped. Not for JTBD-specific force extraction (`interview-insights`), cleaning the raw transcript first (`transcript-cleanup`), or writing the guide that produced the interview (`mom-test-guide`).

Category

Discovery & Research

Output

Use when a cleaned interview transcript needs atomized, citable notes under a chosen framework…

Run

/pm-os:interview-notes

Use when leadership has a gut feeling about a feature and it needs to become a research plan before anyone builds anything — a hunch and company context in, the hunch's hidden assumptions surfaced, and a plan to validate or refute it out. Not for planning a study around an already-defined problem and solution (`research-brief`), framing a decision as a researchable question (`research-decision`), or checking whether old research has decayed (`research-recency`).

Category

Discovery & Research

Output

Use when leadership has a gut feeling about a feature and it needs to become a research plan before anyone…

Run

/pm-os:intuition-to-research

Use when you have user behavior data and want the end-to-end customer experience mapped as ordered stages — touchpoints, time spent, friction and delight, and the emotional state at each — to find where the experience breaks. Every stage grounded in data; inferred emotions flagged. Not for redesigning one flow's friction (`friction-reduce`), the first-run activation path (`onboarding-redesign`), or an empathy map of a single persona (`create-empathy-maps`).

Category

Discovery & Research

Output

Use when you have user behavior data and want the end-to-end customer experience mapped as ordered stages —…

Run

/pm-os:journey-map

Use when multiple coded interviews (JTBD insight sets, one per participant) need to become one clusterable dataset — pushes, pulls, habits, and anxieties coded consistently across the sample so patterns become visible. Also /research Step 3, taking the Step 2 insight sets as input. Not for extracting the four forces from one interview (`interview-insights`), diagnosing one buyer's stuck switch (`diagnose-the-switch`), or writing a quick job story (`jtbd-jobs`).

Category

Discovery & Research

Output

Use when multiple coded interviews (JTBD insight sets, one per participant) need to become one clusterable…

Run

/pm-os:jtbd-forces

Use when you have a single transcript or set of notes and just need job stories written from it — no forces, no intensity tagging, just the situation/motivation/outcome pattern. Not for the full four-forces JTBD extraction (`interview-insights`), a buyer-switch diagnosis (`diagnose-the-switch`), or clustering forces across many interviews (`jtbd-forces`).

Category

Discovery & Research

Output

Use when you have a single transcript or set of notes and just need job stories written from it — no forces,…

Run

/pm-os:jtbd-jobs

Use when a product needs its market redefined through a Jobs-to-be-Done lens — from the traditional product category down to the abstracted job executor and the higher-order job the product actually gets hired for. Works from what you already know about the product, not new interview data — the JTBD Market Definition Canvas. Not for extracting forces from real interviews (`interview-insights`), clustering forces across a sample (`jtbd-forces`), or diagnosing where one specific buyer sits on the switch (`diagnose-the-switch`).

Category

Discovery & Research

Output

Use when a product needs its market redefined through a Jobs-to-be-Done lens — from the traditional product…

Run

/pm-os:jtbd-market-canvas

Use when you need a realistic practice interview transcript to rehearse JTBD technique or demo the analysis skills on — not a real customer, a simulated one, built to surface the four forces of progress naturally without naming them. Not for extracting forces from a real interview (`interview-insights`), or writing the guide to run a real one (`mom-test-guide`).

Category

Discovery & Research

Output

Use when you need a realistic practice interview transcript to rehearse JTBD technique or demo the analysis…

Run

/pm-os:jtbd-transcript-sim

Use when a stated problem needs reframing through De Bono's provocation technique — deliberate disruptions (reversal, exaggeration, distortion, random connection) that break the frame and generate genuinely new options. Not for open-topic idea breadth (`brainstorm-genius`), constraint-driven ideation (`constrained-ideas`), or borrowing structure from an unrelated domain (`metaphor-thinking`).

Category

Ideation & Analytical Thinking

Output

Use when a stated problem needs reframing through De Bono's provocation technique — deliberate disruptions…

Run

/pm-os:lateral-thinking

Use when reviewing your own performance in a past meeting to sound more senior next time — a transcript and your statements in, cited issues across influence/clarity/presence out, each with a specific adjustment and one prioritized focus. Not for designing influence tactics before a meeting (`cialdini`), surfacing attendees' hidden agendas (`hidden-agendas`), or restructuring a piece of writing for clarity (`pyramid-principle`).

Category

Communication, Writing & Influence

Output

Use when reviewing your own performance in a past meeting to sound more senior next time — a transcript and…

Run

/pm-os:leadership-presence

Use when someone wants to design a more satisfying life by finding the *patterns* behind their happiest moments — walk their adult years, capture what genuinely satisfied them, and surface the recurring themes to build more of. Not for career-direction advice from career thoughts (`career-guidance`) or tracking motivation over time (`motivation-journal`).

Category

Coaching, Self-Management & Career

Output

Use when someone wants to design a more satisfying life by finding the *patterns* behind their happiest…

Run

/pm-os:life-design-mapping-exercise

Use when a strategy needs sharpened by envisioning its idealized end state and backcasting to today — a strategy draft or problem, plus the crux behind it, in, a limit state with convergence milestones and an execution plan out. Also `/strategy` Step 2, taking the Step 0 crux and Step 1b strategy draft as input. Not for pushing a variable to its limit to break a fixed-constraint claim (`first-principles`), mapping value generators in the current state (`value-chain`), scanning external trends for new opportunities (`inflection-scan`), or building the strategy document itself (`product-strategy`).

Category

Strategy & Vision

Output

Use when a strategy needs sharpened by envisioning its idealized end state and backcasting to today — a…

Run

/pm-os:limit-strategy

A decision-making framework grounded in Assembly Theory for increasing the likelihood of fortunate outcomes. Use this skill when facing ambiguous choices, formulating experiments, designing strategies, evaluating opportunities, building things meant to persist, or when the user asks about improving their luck, fortune, resilience, or why some ideas, products, or systems thrive while others don't.

Category

Decisions & Prioritization

Output

A decision-making framework grounded in Assembly Theory for increasing the likelihood of fortunate outcomes

Run

/pm-os:luck

Make high-quality product decisions by working through root causes, classifying reversibility, journaling the decision, structuring the problem, synthesizing recommendations, and defining decision rights. Use when making a tradeoff, classifying reversibility, or running a structured decision audit.

Category

Decisions & Prioritization

Output

Make high-quality product decisions by working through root causes, classifying reversibility, journaling the…

Run

/pm-os:decisions

Use when the user says "I need to give a date", "they're asking when it ships", "I have to write the exec update", "I keep putting off status updates", "leadership wants this template filled in", or names any deliverable aimed at execs, GTM, or skip-levels — produces a two-date plan, a finished optics artifact, or a containment setup. Also `/stakeholder` Fast path, the upward-comms counterpart to Step 7's `exec-update-review`. Not for deciding whether a fight is worth capital (`pick-your-battles`), a live incident needing an immediate response (`respond-under-fire`), or prepping for one specific meeting (`prep-the-room`). Source: Shreyas Doshi corpus + The Prince / 48 Laws of Power.

Category

Stakeholders & Politics

Output

Use when the user says "I need to give a date", "they're asking when it ships", "I have to write the exec…

Run

/pm-os:manage-the-upward-channel

Use when a messy business, product or organizational problem needs decomposing into a McKinsey-style issue tree — shape auto-detected as Why (root causes → hypotheses), What (components → sequenced workplan) or How (interventions → ranked options), framed and cruxed before the tree, MECE before the deliverable. Fires on "structure this problem", "build me an issue tree", "find the root cause", "what does producing X require", "decompose this", "why is X happening", "how might we achieve X". Not for a MECE decomposition of an existing list alone (`mece-tree`), root-cause mapping without the hypothesis pipeline (`causal-tree`), converging to one recommendation without a tree (`structure-problem`), or framing the problem before any analysis (`problem-framing-canvas`).

Category

Ideation & Analytical Thinking

Output

Use when a messy business, product or organizational problem needs decomposing into a McKinsey-style issue…

Run

/pm-os:mckinsey-issue-tree

Turn vague intangibles into quantified decision inputs by chaining clarification, decomposition, value-of-information analysis, and small-sample measurement into a single structured workflow. Use when defining KPIs, deciding what to measure, or designing ROI analysis on intangibles.

Category

Metrics & Measurement

Output

Turn vague intangibles into quantified decision inputs by chaining clarification, decomposition,…

Run

/pm-os:measure

Use when a list of items (options, causes, segments, anything) needs a MECE check before you build on it: no overlaps, no gaps, no mixed types, and a rebuilt tree re-checked against all three before it counts. Also /decisions Step 4 and an optional check inside /opportunity, taking whatever list the workflow has already assembled. Not for tracing root causes (`causal-tree`), the full Why/What/How issue-tree pipeline (`mckinsey-issue-tree`), or converging to one recommendation (`structure-problem`).

Category

Decisions & Prioritization

Output

Use when a list of items (options, causes, segments, anything) needs a MECE check before you build on it: no…

Run

/pm-os:mece-tree

Run high-impact meetings — surface hidden agendas before, apply influence principles during, and capture structured summaries after. Use when preparing a meeting agenda, summarizing a meeting, or preparing influence tactics.

Category

Meetings & Facilitation

Output

Run high-impact meetings — surface hidden agendas before, apply influence principles during, and capture…

Run

/pm-os:meeting

Use when you need a full pre-meeting brief before a cross-functional or peer meeting — the decision it's meant to produce, who's in the room and what they want, a time-boxed agenda, and whether a meeting is even the right format. Not for objection-handling prep on a specific adversarial stakeholder meeting (`meeting-prep`), a fast field-read on the people in the room (`prep-the-room`), or the post-meeting summary (`meeting-summary`).

Category

Meetings & Facilitation

Output

Use when you need a full pre-meeting brief before a cross-functional or peer meeting — the decision it's…

Run

/pm-os:meeting-outcomes

Use when you're walking into a genuinely adversarial meeting — stakeholders who are skeptical, tough, possibly unfair — and need to survive their hardest questions before you're in the room. Also /stakeholder Step 5, taking the comms plan and risk register as input. Not for a fast field-read on friendlier room dynamics (`prep-the-room`), the general pre-meeting brief (`meeting-outcomes`), or the difficult-conversation script itself (`difficult-conversation`).

Category

Meetings & Facilitation

Output

Use when you're walking into a genuinely adversarial meeting — stakeholders who are skeptical, tough,…

Run

/pm-os:meeting-prep

Use when you have a meeting transcript or notes and need a structured, skimmable summary in the *IDEAS* format — Insights, Decisions, Engagements, Actions, Summary — with an owner and a date on every action. Also /meeting Step 3. Not for reconstructing a work conversation's flow and open threads (`transcript-insights`), JTBD analysis of a customer interview (`interview-insights`), or extracting structured requirements (`requirements-from-talk`).

Category

Meetings & Facilitation

Output

Use when you have a meeting transcript or notes and need a structured, skimmable summary in the *IDEAS*…

Run

/pm-os:meeting-summary

Use when you're stuck in a stale frame on a problem and want a fresh angle by borrowing structure from an unrelated domain — a *metaphor* that transfers a mechanism, not a mood, into concrete moves the literal frame missed. Not for reframing the problem statement directly (`lateral-thinking`), generating idea volume on a topic (`brainstorm-genius`), or structured analysis toward a recommendation (`structure-problem`).

Category

Ideation & Analytical Thinking

Output

Use when you're stuck in a stale frame on a problem and want a fresh angle by borrowing structure from an…

Run

/pm-os:metaphor-thinking

Use when a metric fell and nobody knows why — proving exceptional variation with an XmR process behaviour chart before investigating, then decomposing drivers, dating the break, segmenting the culprit, and leaving with ranked testable hypotheses. Not for defining the metric and its guardrails before work starts (`success-metric`), instrumenting the events a diagnosis would need (`tracking-schema`), or reviewing a finished experiment's writeup (`feature-writeup`).

Category

Metrics & Measurement

Output

Use when a metric fell and nobody knows why — proving exceptional variation with an XmR process behaviour…

Run

/pm-os:metric-drop

Use when someone states a worry or self-doubt and needs it reframed into something empowering — affirmations and questions that counter the specific negative belief, not generic positivity. Not for a structured strength-discovery interview (`superpowers`), or PM performance coaching across a full situation (`coaching`).

Category

Coaching, Self-Management & Career

Output

Use when someone states a worry or self-doubt and needs it reframed into something empowering — affirmations…

Run

/pm-os:mindset-shift

Use when research has piled up and needs converting into candidate *moats* — ten distinct defensibility strategies scored on time-to-moat, cost, and probability, then narrowed to a three-strategy portfolio with kill criteria and a six-month sequence. Not for diagnosing which of Helmer's 7 Powers the business already holds (`seven-powers`), distilling the loop behind past wins (`flywheel`), or calling match/differentiate on one competitor feature (`parity-vs-differentiation`).

Category

Strategy & Vision

Output

Use when research has piled up and needs converting into candidate *moats* — ten distinct defensibility…

Run

/pm-os:moats

Use when you're about to run a real customer interview and need the discussion guide written first — grounded in *The Mom Test* discipline (past behavior over hypotheticals, no pitching, no leading questions) and Teresa Torres' *Continuous Discovery Habits*. Not for a simulated practice transcript (`jtbd-transcript-sim`), or analyzing a transcript once it exists (`interview-insights`).

Category

Discovery & Research

Output

Use when you're about to run a real customer interview and need the discussion guide written first — grounded…

Run

/pm-os:mom-test-guide

Use when a PM wants to track what actually energizes them at work versus what they think "should" motivate them, entry over entry, to find the real pattern. Takes a new entry, and prior entries if this is a recurring practice. Not for one-time ADHD task motivation (`adhd-motivation`), a life-satisfaction pattern across years (`life-design-mapping-exercise`), or a founder's execution plan (`founder-focus`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants to track what actually energizes them at work versus what they think "should" motivate…

Run

/pm-os:motivation-journal

Use when someone is stuck in two minds about a work change they keep not making — take the role or stay, push back on the exec or go along, run the discovery they promised or keep shipping — and needs their own reasons drawn out rather than advice. Ambivalence in, their change talk and one step they chose out. Not for reframing a single negative belief (`mindset-shift`), generating tactics for a stuck ADHD task (`adhd-motivation`), or tracking what energizes them entry over entry (`motivation-journal`).

Category

Coaching, Self-Management & Career

Output

Use when someone is stuck in two minds about a work change they keep not making — take the role or stay, push…

Run

/pm-os:motivational-interview

Use when a market opportunity needs a Market Requirements Document — TAM, SAM and SOM each carrying a source link or a marked assumption, Porter's Five Forces, buyer behaviour, and the market requirements that fall out of them — built through a question-at-a-time interview rather than a one-shot draft. Not for a feature-level product requirements document (`prd-draft`), redefining the market through Jobs-to-be-Done (`jtbd-market-canvas`), pointing existing company strengths at one industry (`industry-strategy`), or profiling named competitors one by one (`netmba-competitor-analysis`).

Category

PRD, Specs & Documentation

Output

Use when a market opportunity needs a Market Requirements Document — TAM, SAM and SOM each carrying a source…

Run

/pm-os:mrd

Review any product document from 7 cross-functional perspectives — Engineering, Design, Executive, Legal, UX Research, Devil's Advocate, and Customer Voice — with framework-grounded feedback and Socratic questioning. Use when reviewing a PRD, strategy doc, roadmap, or design from multiple stakeholder perspectives.

Category

PRD, Specs & Documentation

Output

Review any product document from 7 cross-functional perspectives — Engineering, Design, Executive, Legal, UX…

Run

/pm-os:pm-review

Use when one named competitor needs a full teardown on Porter's four components — objectives, current strategy, assumptions, capabilities — read forward into their likely moves and the differentiation gaps those moves leave open. Also `/strategy` Step 1a, taking the crux as its focus. Not for calling match/differentiate/leapfrog on a single shipped feature (`parity-vs-differentiation`), writing the positioning statement (`positioning`), or scanning external trends for future shifts (`inflection-scan`).

Category

Strategy & Vision

Output

Use when one named competitor needs a full teardown on Porter's four components — objectives, current…

Run

/pm-os:netmba-competitor-analysis

Use when a broad target needs narrowing into specific, targetable beachhead niches — 10 candidates, each precise enough to state in one sentence combining demographic, interest, geography, and a distinguishing feature, not a broad category restated. Not for the positioning statement once a niche is chosen (`positioning`), or the full competitive strategy (`product-strategy`).

Category

Strategy & Vision

Output

Use when a broad target needs narrowing into specific, targetable beachhead niches — 10 candidates, each…

Run

/pm-os:niche-finder

Use when user interviews need turning into a Now/Next/Later roadmap where every item is grounded in a quote — needs separated from the symptoms and proposed solutions users voice, then placed on the horizon where each becomes *critical*. Not for bridging a vision deck into this quarter's work (`vision-to-quarter`), triaging one week's pile (`weekly-top-3`), or building an opportunity tree from the same interviews (`ost`).

Category

Strategy & Vision

Output

Use when user interviews need turning into a Now/Next/Later roadmap where every item is grounded in a quote —…

Run

/pm-os:now-next-later

Use when you have NPS scores and open-text feedback and want a prioritized set of CX initiatives — detractor and promoter comments split, themed with honest counts and verbatim quotes, then turned into ranked initiatives. Not for general survey open-text (`survey-to-actions`), defining the metric itself (`success-metric`), or the full measurement workflow (`measure`).

Category

Discovery & Research

Output

Use when you have NPS scores and open-text feedback and want a prioritized set of CX initiatives — detractor…

Run

/pm-os:nps-to-cx-plan

Use when you need to anticipate and prepare for stakeholder objections using the MOO (Most Obvious Objection) approach — a proposed initiative in, ranked objections per stakeholder group with counter-moves and pre-wire scripts out. Not for per-person motivation and fear mapping (`hidden-agendas`), auditing a specific PRD for circulation risk (`stakeholder-risk`), or generating training-only flawed solutions (`solution-flaws`).

Category

Stakeholders & Politics

Output

Use when you need to anticipate and prepare for stakeholder objections using the MOO (Most Obvious Objection)…

Run

/pm-os:objection-scan

Use when a topic needs playful, absurd, or humorous questions to loosen rigid thinking before serious ideation — unexpected comparisons and playful twists, not high-leverage strategic questions. Not for serious, high-leverage strategic questions (`good-question-brainstormer`), or provocation-based problem reframing (`lateral-thinking`).

Category

Ideation & Analytical Thinking

Output

Use when a topic needs playful, absurd, or humorous questions to loosen rigid thinking before serious…

Run

/pm-os:offbeat-questions

Use when a product's onboarding needs redesigning around the user's *transformation* — who they are before, who they become after, and the smallest step that makes them feel the change — with every current step audited as serving the user or serving the product. Fires on "our onboarding leaks users" or "redesign our first-run experience." Not for locating the activation metric (`aha-moment`), mapping the full journey (`journey-map`), or general flow friction (`friction-reduce`).

Category

Design & Prototyping

Output

Use when a product's onboarding needs redesigning around the user's *transformation* — who they are before,…

Run

/pm-os:onboarding-redesign

Map product opportunities systematically — from structured OST intake through opportunity tree construction, optional MECE refinement, to selecting one prioritized target opportunity. Use when mapping product opportunities, building an opportunity solution tree, or selecting initiatives to pursue.

Category

Decisions & Prioritization

Output

Map product opportunities systematically — from structured OST intake through opportunity tree construction,…

Run

/pm-os:opportunity

Use when interview material needs structuring into a Teresa Torres Opportunity Solution Tree — opportunities extracted quote by quote, assigned to *moments in time*, deduplicated, and checked with the sibling-distinctness and parent-child tests. Also `/opportunity` Step 2. Not for eliciting the inputs first (`ost-intake`), selecting one target from the finished tree (`ost-prioritize`), or MECE-checking a flat list of siblings (`mece-tree`).

Category

Decisions & Prioritization

Output

Use when interview material needs structuring into a Teresa Torres Opportunity Solution Tree — opportunities…

Run

/pm-os:ost

Use when an Opportunity Solution Tree is about to be built and its four inputs need eliciting and normalizing first — outcome, journey nodes as *moments in time*, interview material, constraints — parsing whatever the user already provided before asking anything. Also `/opportunity` Step 1. Not for building the tree from those inputs (`ost`), selecting a target from a built tree (`ost-prioritize`), or scoping a research study (`research-brief`).

Category

Decisions & Prioritization

Output

Use when an Opportunity Solution Tree is about to be built and its four inputs need eliciting and normalizing…

Run

/pm-os:ost-intake

Use when a mapped opportunity space needs narrowing to one target to explore next — Teresa Torres's four factors applied as *compare-and-contrast* reasoning, never a weighted score, from a tree, a flat list, or an OST JSON. Also `/opportunity` Step 3. Not for building the tree first (`ost`), force-ranking a requirements list into P0/P1/P2 (`p0-p1-p2`), or placing one request on a priority×effort 2×2 (`tradeoff-matrix`).

Category

Decisions & Prioritization

Output

Use when a mapped opportunity space needs narrowing to one target to explore next — Teresa Torres's four…

Run

/pm-os:ost-prioritize

Use when a requirements list has everything marked "must-have" and the PM needs it forced into P0/P1/P2 — P0 meaning *launch blocker*, tested, not claimed. Fires on "prioritize these requirements," "what's actually P0 here," or pre-launch scope fights. Not for pricing one incoming scope request (`scope-defense`) or ranking discovery opportunities (`ost-prioritize`).

Category

Decisions & Prioritization

Output

Use when a requirements list has everything marked "must-have" and the PM needs it forced into P0/P1/P2 — P0…

Run

/pm-os:p0-p1-p2

Use when someone says "competitor X just shipped Y — we need it" and the PM must call it: match, differentiate, leapfrog, or skip. The hinge is whether the feature is *table stakes* — proven by lost deals and churn, not by its existence. Not for a full competitor teardown (`netmba-competitor-analysis`), building moats (`moats`), or counter-positioning strategy (`disruption-what-ifs`).

Category

Strategy & Vision

Output

Use when someone says "competitor X just shipped Y — we need it" and the PM must call it: match,…

Run

/pm-os:parity-vs-differentiation

Use when a specific decision or topic needs assessing for how one particular user segment would react — reaction, fit, adoption barriers, and communication angle, from a third-person analytical lens. Optionally takes a `create-empathy-maps` output as input. Not for interactively play-acting one customer's first-person voice across arbitrary questions (`icp-roleplay`), or general audience psychology not tied to one decision (`create-empathy-maps`).

Category

Discovery & Research

Output

Use when a specific decision or topic needs assessing for how one particular user segment would react —…

Run

/pm-os:persona-lens-decision

Use when the user says "should I fight this", "should I escalate", "should I push back on this decision", "this process/strategy is broken and I want to fix it", "I want to go over someone's head", "how do I take this on", or is contemplating any deliberate contest: escalation, open disagreement with a decision, a move against a rival or an exec, or a campaign to change something — produces a GO/NO-GO with capital budget; if GO, a designed move or consolidated strike plan. Also `/stakeholder` Fast path. Not for a live incident needing an immediate response (`respond-under-fire`), managing what's promised upward instead of fighting (`manage-the-upward-channel`), or prepping for one specific meeting (`prep-the-room`). Source: Shreyas Doshi corpus + The Prince / 48 Laws of Power.

Category

Stakeholders & Politics

Output

Use when the user says "should I fight this", "should I escalate", "should I push back on this decision",…

Run

/pm-os:pick-your-battles

Use when a company's growth options need the Pioneer-Migrator-Settler map — the current *portfolio* placed on it first, the imbalance read, then growth ideas generated to fill the gap. Fires on "are we too red-ocean," "PMS map," or Blue Ocean portfolio reviews. Not for the growth model and its constraint (`growth-strategy`), counter-positioning ideation (`disruption-what-ifs`), or SWOT-derived moves (`swot-moves`).

Category

Strategy & Vision

Output

Use when a company's growth options need the Pioneer-Migrator-Settler map — the current *portfolio* placed on…

Run

/pm-os:pioneer-migrator-settler

PM coaching grounded in what designers and engineers actually say about their best product managers. Run situation retrospectives, reveal how your team sees you, stress-test yourself in adversarial roleplay, audit past decisions, and surface the blind spots you can't see from the inside. Use when the user wants to debrief a situation, audit a decision, do an adversarial roleplay, or scan blind spots.

Category

Coaching, Self-Management & Career

Output

PM coaching grounded in what designers and engineers actually say about their best product managers

Run

/pm-os:coaching

Use when preparing for a PM interview — a resume, JD, interview context (type, focus, duration), and specific concerns in, a tailored set of anticipated questions out, each with why it's asked, what's being evaluated, and an answer framework. Not for drafting the STAR stories those answers rely on (`star-stories`), tailoring the resume itself (`resume`), or broader career-direction guidance (`career-guidance`).

Category

Coaching, Self-Management & Career

Output

Use when preparing for a PM interview — a resume, JD, interview context (type, focus, duration), and specific…

Run

/pm-os:pm-interview

Use when the user wants to create, author, or scaffold their own custom agent or subagent, says 'make my own agent', 'build me an agent', or types /agent-builder. Not for finding an existing PM skill that already does the job (`pm-os-skill`) or touring what the system already has (`pm-help`).

Category

PM OS System & Operations

Output

Use when the user wants to create, author, or scaffold their own custom agent or subagent, says 'make my own…

Run

/pm-os:pm-os-agent-builder

Use when the user wants to save a decision, risk, or PM update to project memory, or says '/capture-memory' or 'remember this'. Not for the slow one-question-a-day accretion loop (`pm-os-daily-drip`), one-time bulk import from another AI assistant (`import-ai-memory`), or managing the project folder itself rather than what's written into it (`pm-os-project`).

Category

PM OS System & Operations

Output

Use when the user wants to save a decision, risk, or PM update to project memory, or says '/capture-memory'…

Run

/pm-os:pm-os-capture-memory

Use when the user wants to run the opt-in daily context question loop, answer a pending question, or says '/daily-drip'. Not for explicit one-off capture of a decision or update (`pm-os-capture-memory`) or one-time bulk import from another AI assistant (`import-ai-memory`).

Category

PM OS System & Operations

Output

Use when the user wants to run the opt-in daily context question loop, answer a pending question, or says…

Run

/pm-os:pm-os-daily-drip

Use when the user wants to send feedback to the pm-os creator, share what's broken, or says '/feedback'. Not for a structured testimonial written for marketing use (`pm-os-testimonial`) or coaching follow-up on a recurring behavior pattern (`good-pm-bad-pm`).

Category

PM OS System & Operations

Output

Use when the user wants to send feedback to the pm-os creator, share what's broken, or says '/feedback'

Run

/pm-os:pm-os-feedback

Use when the user wants to find a framework, asks 'what framework should I use', or says '/framework'. Not for finding a PM skill for a specific task (`pm-os-skill`) or a full system tour (`pm-help`).

Category

PM OS System & Operations

Output

Use when the user wants to find a framework, asks 'what framework should I use', or says '/framework'

Run

/pm-os:pm-os-framework

Use when the user wants to create, list, switch, activate, or archive a PM OS project folder, or says '/project'. Not for routing orphaned files into the folders this creates (`pm-os-tidy`) or capturing an event into the project once it exists (`pm-os-capture-memory`).

Category

PM OS System & Operations

Output

Use when the user wants to create, list, switch, activate, or archive a PM OS project folder, or says…

Run

/pm-os:pm-os-project

Use when the user wants to find the right PM skill for a task, browse skills by topic, or says '/skill'. Not for finding a framework (`pm-os-framework`), a full system tour (`pm-help`), or browsing every skill visually (`skill-browser`).

Category

PM OS System & Operations

Output

Use when the user wants to find the right PM skill for a task, browse skills by topic, or says '/skill'

Run

/pm-os:pm-os-skill

Use when the user wants to onboard, set up their PM OS Context files, or says '/start'. Not for a one-time paste-back import from another AI assistant standalone (`import-ai-memory` — this skill invokes it as one branch) or merging a version upgrade into an existing workspace (`pm-os-upgrade`).

Category

PM OS System & Operations

Output

Use when the user wants to onboard, set up their PM OS Context files, or says '/start'

Run

/pm-os:pm-os-start

Use when the user wants to give a testimonial about pm-os or says '/testimonial'. Not for free-form feedback or bug reports (`pm-os-feedback`), or capturing an internal decision or update into project memory (`pm-os-capture-memory`).

Category

PM OS System & Operations

Output

Use when the user wants to give a testimonial about pm-os or says '/testimonial'

Run

/pm-os:pm-os-testimonial

Use when the user wants to sync Context with Work, tidy up orphaned artifacts, route files into project folders, or says '/tidy' or 'clean up'. Not for creating or archiving the project folders themselves (`pm-os-project`) or capturing a specific decision or update into memory (`pm-os-capture-memory`).

Category

PM OS System & Operations

Output

Use when the user wants to sync Context with Work, tidy up orphaned artifacts, route files into project…

Run

/pm-os:pm-os-tidy

Use when the user wants to upgrade an existing pm-os workspace to a newer version, or says '/upgrade' or '/pm-upgrade'. Not for a fresh install (`pm-os-start`) or creating/managing project folders after the upgrade (`pm-os-project`).

Category

PM OS System & Operations

Output

Use when the user wants to upgrade an existing pm-os workspace to a newer version, or says '/upgrade' or…

Run

/pm-os:pm-os-upgrade

Use when the user wants to orient on current state, asks 'where am I', wants a recommended next action, or says '/status' or '/pm-status'. Not for a full command tour (`pm-help`) or listing/switching project folders (`pm-os-project`).

Category

PM OS System & Operations

Output

Use when the user wants to orient on current state, asks 'where am I', wants a recommended next action, or…

Run

/pm-os:pm-status

Use when a PM's context (goals, success metrics, business requirements) needs translating into concrete design constraints — especially the constraints a PM assumes but never states. Not for the single-screen kickoff sketch (`app-design-foundation`), full flow sketches (`wireframe-sketch`), or reverse-engineering requirements from a design that already exists (`requirements-from-design`).

Category

Design & Prototyping

Output

Use when a PM's context (goals, success metrics, business requirements) needs translating into concrete…

Run

/pm-os:pm-to-design

Use when PM OS's own voice is the thing being worked on and you invoke this by name — grading a response against the PM Thinking Partner contract ("does this sound like a CPO?", "audit this for voice"), rewriting text into that voice, or applying plain sentence construction ("simplify this", "STE100 rewrite", "make this unambiguous"). The contract in `AGENTS.md` is already always-on, so this fires by request rather than on its own. Not for restructuring arbitrary content for readability (`skimmable-writing`), testing whether a claim is clever rather than true (`de-clever`), or leading a buried document with its conclusion (`pyramid-principle`).

Category

Communication, Writing & Influence

Output

Use when PM OS's own voice is the thing being worked on and you invoke this by name — grading a response…

Run

/pm-os:pm-voice

Use when a product needs a positioning statement and value proposition — from a competitive analysis, a stated value proposition, or both — a Geoffrey Moore–style statement that's falsifiable and outcome-led, not feature-led. Not for the underlying competitive analysis itself (`netmba-competitor-analysis`), the full structured strategy document (`product-strategy`), or the complete strategy workflow (`/strategy`).

Category

Strategy & Vision

Output

Use when a product needs a positioning statement and value proposition — from a competitive analysis, a…

Run

/pm-os:positioning

Use when you need to know who actually holds power in a stakeholder system before acting — a decision or initiative and the people involved in, a ranked power map out (formal authority, informal influence, incentives, relationships, and influence pathways). Also `/stakeholder` Step 1, the entry point before Step 2's `stakeholder-map` grid. Not for prepping the adversarial meeting itself (`meeting-prep`), plotting stakeholders on a Power-Interest grid (`stakeholder-map`), or surfacing what people want rather than how much say they have (`hidden-agendas`).

Category

Stakeholders & Politics

Output

Use when you need to know who actually holds power in a stakeholder system before acting — a decision or…

Run

/pm-os:power-map

Use when retained users need a transition plan from a low-frequency use case to a high-frequency one — the behavior gap named, adoption barriers identified, and a phased rollout with onboarding and metrics. Not for winning back users who already left (`churn-reduction`), or the broader growth-loop diagnosis (`growth-strategy`).

Category

Metrics & Measurement

Output

Use when retained users need a transition plan from a low-frequency use case to a high-frequency one — the…

Run

/pm-os:power-user-shift

Build a complete PRD end-to-end — from raw inputs (transcript, design, research) through use cases, user stories, and acceptance criteria to a drafted PRD reviewed by 9 specialised reviewers (2 PRD-specific personas + 7 cross-functional). Use when you need a full PRD, not just a one-shot draft.

Category

PRD, Specs & Documentation

Output

Build a complete PRD end-to-end — from raw inputs (transcript, design, research) through use cases, user…

Run

/pm-os:prd

Use when structured requirements need to become a full PRD in one pass — the format picked from the `📄 Templates/` library by stage, audience and detail level, every section filled, and every assumption you had to make listed as a follow-up question instead of buried. Covers enterprise scope (security, compliance, performance, technical requirements) when the product is regulated or enterprise-sold. Also /prd Step 6, assembling Steps 1–5 into the draft. Not for the full multi-step PRD pipeline with reviewers (`prd`), a market-level requirements document (`mrd`), a working-backwards launch announcement (`press-release`), or the technical architecture brief that follows the PRD (`tech-arch-brief`).

Category

PRD, Specs & Documentation

Output

Use when structured requirements need to become a full PRD in one pass — the format picked from the `📄…

Run

/pm-os:prd-draft

Use when the user says "prep me for this meeting", "I'm walking into a conversation with X", "I have a review with my VP in an hour", "how do I pitch this to the team", "help me get ready for this 1:1", or names an upcoming interaction with a stakeholder, exec, engineer, or designer that carries tension or stakes — produces a read on the actors, a stance, opening moves, and what not to do. Also `/stakeholder` Fast path. Not for a planned, already-identified hard conversation needing a full script (`difficult-conversation`), deciding whether the underlying issue is worth fighting (`pick-your-battles`), or a reactive incident that already happened (`respond-under-fire`). Source: Shreyas Doshi org-navigation corpus + The Prince / 48 Laws of Power.

Category

Stakeholders & Politics

Output

Use when the user says "prep me for this meeting", "I'm walking into a conversation with X", "I have a review…

Run

/pm-os:prep-the-room

Use when you need the full narrative prose for a persuasive presentation — a three-arc story built from raw material (transcripts, notes, a brief) that ends in a specific buy-in ask. Not for planning the deck for an audience first (`strategic-deck`), a slide-by-slide outline (`problem-deck`), or scripting a product demo (`demo-narrative`).

Category

Communication, Writing & Influence

Output

Use when you need the full narrative prose for a persuasive presentation — a three-arc story built from raw…

Run

/pm-os:presentation-narrative

Use when a product needs an Amazon-style working-backwards PR/FAQ — the release written as though it ships today with a customer quote, a leader quote and one hard proof point, then the external and internal FAQs that answer what press, customers and leadership will actually ask. Every unknown is left as a marked placeholder rather than invented. Not for a product requirements document (`prd-draft`), an outage or incident announcement (`crisis-comms`), a routine stakeholder communication plan (`comms-plan`), or the positioning statement the release leans on (`positioning`).

Category

PRD, Specs & Documentation

Output

Use when a product needs an Amazon-style working-backwards PR/FAQ — the release written as though it ships…

Run

/pm-os:press-release

Use when you need to turn a problem and a recommendation into a slide-by-slide storyline — question-title slides where each answers the one before, building to the ask (Minto/Pyramid + What/So What/Now What). Not for arranging content you already have (`what-so-what-deck`), writing the narrative prose (`presentation-narrative`), or planning depth and tone for an audience (`strategic-deck`).

Category

Communication, Writing & Influence

Output

Use when you need to turn a problem and a recommendation into a slide-by-slide storyline — question-title…

Run

/pm-os:problem-deck

Use when a team has locked onto a solution and needs pulling back to the problem — the *solution jump* diagnosed, the embedded assumptions surfaced and challenged, alternative framings generated, and validation research proposed before design starts. Not for turning a vague stakeholder brief into a problem statement (`brief-to-problem`), framing from a workshop transcript (`problem-framing-canvas`), or finding the pivotal obstacle in a strategy problem (`find-the-strategic-crux`).

Category

Decisions & Prioritization

Output

Use when a team has locked onto a solution and needs pulling back to the problem — the *solution jump*…

Run

/pm-os:problem-first

Use when you have a session transcript, interview, or conversation and need the core problem framed from the person experiencing it — a first-person narrative (I am / trying to / but / because / feels), the constraints around them, and one problem statement stakeholders can rally behind. Also invoked by `mckinsey-issue-tree` Phase 3 to frame before building the tree. Not for turning a vague stakeholder brief into a problem statement (`brief-to-problem`), building the issue tree itself (`mckinsey-issue-tree`), or converging to a recommendation (`structure-problem`).

Category

Discovery & Research

Output

Use when you have a session transcript, interview, or conversation and need the core problem framed from the…

Run

/pm-os:problem-framing-canvas

Use when a product strategy's hidden bets need writing down as falsifiable claims — a core problem, product description, and target user in, "We believe that…" statements across desirability, feasibility, viability, and usability out, each with its impact if wrong. Also `/assumptions` Step 1. Not for ranking which of them to test first (`risky-assumptions`), finding the cheapest signal that would settle one (`work-backwards`), or turning a problem into a single measurable hypothesis (`product-hypothesis`).

Category

Discovery & Research

Output

Use when a product strategy's hidden bets need writing down as falsifiable claims — a core problem, product…

Run

/pm-os:product-assumptions

Use when a problem or opportunity area needs *product concepts* — each stated as problem + mechanism + differentiation, with technology nouns banned so the concept can't hide behind a buzzword. Fires on "product ideas for X," "what could we build here." Not for topic ideation (`brainstorm-genius`), ideation inside hard constraints (`constrained-ideas`), or question-driven startup ideation (`startup-ideas`).

Category

Ideation & Analytical Thinking

Output

Use when a problem or opportunity area needs *product concepts* — each stated as problem + mechanism +…

Run

/pm-os:product-brainstorm

Use when evaluating an existing product design or mockup — screenshots plus context (goals, users, metrics) in, a structured critique out across issues, optimization opportunities, alternatives, and pros/cons, each claim tied to a stated goal or named UX principle. Not for generating new design ideas from a blank challenge (`affordances-signifiers`), correcting UX vocabulary in written requirements (`ux-terminology`), or reviewing a written document (`/review`).

Category

Design & Prototyping

Output

Use when evaluating an existing product design or mockup — screenshots plus context (goals, users, metrics)…

Run

/pm-os:product-design-analyzer

Use when a product problem needs to become a testable hypothesis with a measurable success clause — a problem in, a falsifiable "we believe / we'll know" statement out with a stated metric and threshold. Also `/research` Step 4, taking Step 3's clustered forces as input. Not for designing the experiment that tests the hypothesis (`experiment-design`), clustering the JTBD forces that feed it (`jtbd-forces`), or scoping a research study before any hypothesis exists (`research-brief`).

Category

Discovery & Research

Output

Use when a product problem needs to become a testable hypothesis with a measurable success clause — a problem…

Run

/pm-os:product-hypothesis

Use when product context needs shaping into a structured strategy document through the seven-part template — Objective, Users, *Superpowers*, Vision, Pillars, Impact, Roadmap — where each section must build on the previous ones. Also /strategy Step 1b, taking the crux and competitor analysis as input. Not for diagnosing the strategy problem itself (`strategy-kernel`), backcasting from an end state (`limit-strategy`), or the full strategy workflow (`/strategy`).

Category

Strategy & Vision

Output

Use when product context needs shaping into a structured strategy document through the seven-part template —…

Run

/pm-os:product-strategy

Use when a project brief needs turning into a schedule with a *critical path* — tasks, estimates, milestones, risks, and the ambiguities that will blow the timeline if left unasked. Fires on "plan this project" or "build a timeline from this brief." Not for organizing an existing task list (`task-list-to-plan`), sequencing a to-do dump (`task-sequence`), or prioritizing feedback into a release (`v2-plan`).

Category

Planning & Execution

Output

Use when a project brief needs turning into a schedule with a *critical path* — tasks, estimates, milestones,…

Run

/pm-os:project-plan

Use when research exists but no persona does — interviews, analytics, market data, and demographics in, one proto-persona out where every line is tagged *evidence* or *assumption* with a source, carries a confidence rating, and the unknowns become prioritized probing questions. Not for upgrading an existing shallow persona into tasks and constraints (`user-profile`), mapping one audience's emotions around a topic (`create-empathy-maps`), or play-acting a customer's first-person voice (`icp-roleplay`).

Category

Discovery & Research

Output

Use when research exists but no persona does — interviews, analytics, market data, and demographics in, one…

Run

/pm-os:proto-persona

Use when a piece of writing buries its conclusion under detail and needs Minto's pyramid structure — text in, the key message first, then 2-3 supporting arguments with their strongest evidence, out. Not for designing influence tactics for a specific person (`cialdini`), reviewing your own spoken delivery (`leadership-presence`), or signposting a message so it survives a skim (`skimmable-writing`).

Category

Communication, Writing & Influence

Output

Use when a piece of writing buries its conclusion under detail and needs Minto's pyramid structure — text in,…

Run

/pm-os:pyramid-principle

Use when preparing for or running a product refinement session — a timeboxed agenda, discussion questions, and a closing statement built around understanding the problem deeply, not estimating the solution. Not for turning a raw task list into a sequence (`task-sequence`), or a general meeting prep brief (`meeting-prep`).

Category

Meetings & Facilitation

Output

Use when preparing for or running a product refinement session — a timeboxed agenda, discussion questions,…

Run

/pm-os:refinement-session

Use when a small set of strategic components — a flywheel, an initiative sequence, a set of bets — needs ordering into a self-reinforcing loop, each component making the next easier, with the last strengthening the first. Not for organizing a chaotic to-do dump (`task-sequence`), or discovering the flywheel's components in the first place (`flywheel`).

Category

Decisions & Prioritization

Output

Use when a small set of strategic components — a flywheel, an initiative sequence, a set of bets — needs…

Run

/pm-os:reinforcing-sequence

Use when you have wireframes, mockups, or a Figma export and need requirements reverse-engineered from what's actually shown — every visible element, layout, and interaction turned into a requirement with acceptance criteria a junior engineer could build from. Also /prd Step 1b, one of three input modes. Not for a transcript instead of a design (`requirements-from-talk`), Cockburn-style use cases (`use-cases`), or UI acceptance criteria on stories that already exist (`ui-acceptance-criteria`).

Category

PRD, Specs & Documentation

Output

Use when you have wireframes, mockups, or a Figma export and need requirements reverse-engineered from what's…

Run

/pm-os:requirements-from-design

Use when you have a meeting transcript or interview notes and need a structured requirements list extracted from it — organized MECE, each requirement attributed to who said it, with the gaps and contradictions named rather than smoothed over. Also /prd Step 1a, one of three input modes. Not for a design asset instead of a transcript (`requirements-from-design`), reconciling requirements that already conflict (`requirements-reconcile`), or turning requirements into Cockburn use cases (`use-cases`).

Category

PRD, Specs & Documentation

Output

Use when you have a meeting transcript or interview notes and need a structured requirements list extracted…

Run

/pm-os:requirements-from-talk

Use when stakeholders have handed in *conflicting* requirements and the PM needs one unified list — "sales wants X, engineering wants not-X," multi-stakeholder PRD input, or /prd Step 2. Many conflicts dissolve once each side's underlying goal is named; the rest get decided, not averaged. Not for scoring options on a matrix (`tradeoff-matrix`) or deciding whether to fight at all (`pick-your-battles`).

Category

PRD, Specs & Documentation

Output

Use when stakeholders have handed in *conflicting* requirements and the PM needs one unified list — "sales…

Run

/pm-os:requirements-reconcile

Use when a problem and a proposed solution are already defined and need to become an actionable research brief — background, scope, timeline, methods, deliverables, and expected outcomes, tied to what's actually risky about the solution. Not for framing a decision as a researchable question first (`research-decision`), turning an unvalidated leadership hunch into a research plan (`intuition-to-research`), or synthesizing findings once research is done (`research-synthesis`).

Category

Discovery & Research

Output

Use when a problem and a proposed solution are already defined and need to become an actionable research…

Run

/pm-os:research-brief

Use when you have a decision to make and want to frame it as a specific, researchable *decision question* — sized by how reversible and high-impact the call is — so research is attached to a decision instead of run for its own sake. Not for planning the study itself (`research-brief`), turning a hunch into a research plan (`intuition-to-research`), or synthesizing findings into insights (`research-synthesis`).

Category

Discovery & Research

Output

Use when you have a decision to make and want to frame it as a specific, researchable *decision question* —…

Run

/pm-os:research-decision

Use when a PM is about to reuse old research and needs to know what has *decayed* — "we did interviews 18 months ago, can we still trust them?", pre-kickoff checks on an old study, or challenges to a decision resting on aging findings. Not for synthesizing research (`research-synthesis`) or planning new research from a hunch (`intuition-to-research`).

Category

Discovery & Research

Output

Use when a PM is about to reuse old research and needs to know what has *decayed* — "we did interviews 18…

Run

/pm-os:research-recency

Use when a PM has *scattered* research inputs — interview notes, support tickets, NPS comments, analytics, Slack anecdotes — and needs them merged into a few evidenced insights that answer specific product questions. Also fires on "what is all this feedback telling us?". Not for a single interview transcript (`interview-insights`), one conversation (`transcript-insights`), or survey-only data (`survey-to-actions`).

Category

Discovery & Research

Output

Use when a PM has *scattered* research inputs — interview notes, support tickets, NPS comments, analytics,…

Run

/pm-os:research-synthesis

Transform raw interview transcripts into a tested feature hypothesis — from transcript cleanup through JTBD extraction, clustering, hypothesis formation, and experiment design. Use when synthesizing user interviews, deriving jobs-to-be-done, or turning research into feature hypotheses.

Category

PRD, Specs & Documentation

Output

Transform raw interview transcripts into a tested feature hypothesis — from transcript cleanup through JTBD…

Run

/pm-os:research

Use when the user says "this just happened", "X just told me...", "the date slipped", "I just got bad feedback", "someone is being difficult", "I'm being baited", or describes any live workplace incident needing a response — a slipped date, a missed dependency, an elaborate excuse, a snarky or passive-aggressive comment, negative stakeholder feedback, brewing drama, or a colleague who feels steamrolled — produces a categorized incident, an immediate response with exact scripts, and a forward-motion next step. Also `/stakeholder` Fast path, the reactive counterpart to Step 6's `difficult-conversation`. Not for a planned conversation prepared ahead of time (`difficult-conversation`), deciding whether to escalate a pattern rather than react to one moment (`pick-your-battles`), or a public outage needing org-wide comms (`crisis-comms`). Source: Shreyas Doshi corpus + 48 Laws of Power.

Category

Stakeholders & Politics

Output

Use when the user says "this just happened", "X just told me...", "the date slipped", "I just got bad…

Run

/pm-os:respond-under-fire

Use when tailoring an existing resume to a specific job description — a resume, a JD, and a situation (perfect fit / stretch role / career pivot) in, a JD-matched resume out with every bullet rewritten in the JD's language and grounded in real experience. Not for building the STAR stories behind resume bullets (`star-stories`), preparing spoken interview answers (`pm-interview`), or broader career-direction guidance (`career-guidance`).

Category

Coaching, Self-Management & Career

Output

Use when tailoring an existing resume to a specific job description — a resume, a JD, and a situation…

Run

/pm-os:resume

Use when placing a PM's resume on the career ladder and naming the skill gaps to the next level — a resume in, the current level with reasoning, the next level's requirements, and the specific named skill gaps out. Not for the actual learning plan to close a named gap (`skill-mastery`), broader career-direction guidance beyond a ladder placement (`career-guidance`), or tailoring the resume itself (`resume`).

Category

Coaching, Self-Management & Career

Output

Use when placing a PM's resume on the career ladder and naming the skill gaps to the next level — a resume…

Run

/pm-os:resume-skill-gaps

Use when a list of assumptions needs ranking and a 1–5 scoring rubric would just produce ties — *pairwise* comparisons under QuickSort force one pick per pair, ending in a full order and the top 1–3 to test first. Also `/assumptions` Step 2. Not for generating the assumptions being ranked (`product-assumptions`), finding the cheapest signal that settles one (`work-backwards`), or ranking discovery opportunities rather than risks (`ost-prioritize`).

Category

Discovery & Research

Output

Use when a list of assumptions needs ranking and a 1–5 scoring rubric would just produce ties — *pairwise*…

Run

/pm-os:risky-assumptions

Use when five observations are all the budget or the calendar allows and the median still has to be bounded — a 93.75% confidence interval from the smallest and largest of five *random* samples, with the randomness check that makes the number real. Also /measure Step 5, sampling the uncertainties the value-of-information pass marked worth measuring. Not for estimating a quantity with no samples available at all (`fermi-decomposition`), deciding whether the sample is worth collecting (`value-of-information`), or powering an experiment for significance (`experiment-design`).

Category

Metrics & Measurement

Output

Use when five observations are all the budget or the calendar allows and the median still has to be bounded —…

Run

/pm-os:rule-of-five

Use when a position has to hold up under someone trying to break it — building an argument from premises to conclusion, auditing an existing one for validity and fallacies, drafting an argumentative essay, or steelmanning both sides of a contested call — against Weston's *Rulebook*, with the conclusion held to exactly what the reasons establish. Not for ordering a document so its conclusion leads (`pyramid-principle`), testing whether a framing is clever rather than true (`de-clever`), or decomposing a problem into MECE branches (`mece-tree`).

Category

Communication, Writing & Influence

Output

Use when a position has to hold up under someone trying to break it — building an argument from premises to…

Run

/pm-os:rulebook-for-arguments

Use when a team is stuck reacting and needs a plan built from four views of the situation — current state, the projected *mess* if nothing changes, the ideal future, and what of that ideal can start today. Fires on "where is this heading," "help me plan out of this situation," or adaptive-planning requests. Not for disruption ideation (`disruption-what-ifs`), SWOT move generation (`swot-moves`), or a delivery plan from a task list (`project-plan`).

Category

Planning & Execution

Output

Use when a team is stuck reacting and needs a plan built from four views of the situation — current state,…

Run

/pm-os:scenario-plan

Use when a mid-project "can we just add one more thing" request lands and the PM needs its *cost* made visible before anyone says yes — quantified effort, what it displaces, and decision options with named consequences. Not for prioritizing a whole requirements list (`p0-p1-p2`), visualizing agreed tradeoffs (`tradeoff-matrix`), or deciding whether to fight the requester (`pick-your-battles`).

Category

Decisions & Prioritization

Output

Use when a mid-project "can we just add one more thing" request lands and the PM needs its *cost* made…

Run

/pm-os:scope-defense

Use when a business needs auditing against Hamilton Helmer's 7 Powers — each power scored present, emerging, or absent on falsifiable tests, then acquisition playbooks written for the ones missing. Not for generating new defensibility strategies from a research bundle (`moats`), tearing down one named competitor (`netmba-competitor-analysis`), or distilling the loop behind past wins (`flywheel`).

Category

Strategy & Vision

Output

Use when a business needs auditing against Hamilton Helmer's 7 Powers — each power scored present, emerging,…

Run

/pm-os:seven-powers

Use when a vague feature idea needs shaping into a Shape Up pitch — an appetite set as a constraint rather than an estimate, boundaries drawn, rabbit holes and no-gos named, and the constraint language an executive will actually commit to before the cycle starts. Not for the collaborative shaping methodology with slices and breadboards (`shaping`), holding scope once it is already under pressure (`scope-defense`), finding the application archetype before shaping (`eng-shape`), or a full requirements document (`prd-draft`).

Category

PRD, Specs & Documentation

Output

Use when a vague feature idea needs shaping into a Shape Up pitch — an appetite set as a constraint rather…

Run

/pm-os:shape-up

Use this methodology when collaboratively shaping a solution with the user - iterating on problem definition (requirements) and solution options (shapes).

Category

PRD, Specs & Documentation

Output

Use this methodology when collaboratively shaping a solution with the user - iterating on problem definition…

Run

/pm-os:shaping

Use when a PM wants a debrief on one real situation they just came out of — a meeting, conflict, decision, launch — their account of what happened in, their behavior mapped against the PM excellence clusters out, with strengths named, gaps named with their cost, and one concrete change to make next time. `/coaching` mode 1. Not for auditing one decision's process (`decision-audit`), simulating how the team perceives them (`team-perspective-reveal`), or a systematic audit across all clusters (`blind-spot-scan`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants a debrief on one real situation they just came out of — a meeting, conflict, decision,…

Run

/pm-os:situation-retrospective

Use when the user wants to browse, discover, or get an overview of available skills, or asks "what skills are available", "show me everything I can do", "what can PM OS do", "list all skills", or "browse the skill library". Not for matching one task to one skill (`pm-os-skill`) or a live, personalized command tour (`pm-help`).

Category

PM OS System & Operations

Output

Use when the user wants to browse, discover, or get an overview of available skills, or asks "what skills are…

Run

/pm-os:skill-browser

Use when a vague learning goal needs a falsifiable mastery plan — a skill, a time horizon, and a weekly budget in, a deconstructed sub-skill map, an 80/20 sprint plan, an accountability contract, and a tracked scoreboard out, via the DSSS (Deconstruction, Selection, Sequencing, Stakes) framework. Not for a Socratic tutoring session on a topic (`ai-tutor`), general adaptive PM performance coaching (`coaching`), or broader career-direction guidance (`career-guidance`).

Category

Coaching, Self-Management & Career

Output

Use when a vague learning goal needs a falsifiable mastery plan — a skill, a time horizon, and a weekly…

Run

/pm-os:skill-mastery

Use when a message is written but won't survive a skim — rebuilt with the point and its stakes in the first line, an explicit *signpost* opening every paragraph so the reader always knows what a passage is doing, and 20–40% of the words cut. Not for reordering a buried argument into Minto's key-message-first hierarchy (`pyramid-principle`), a progress update's fixed wins/risks/asks format (`status-update`), or an exec-lens critique of a draft before you present it (`exec-update-review`).

Category

Communication, Writing & Influence

Output

Use when a message is written but won't survive a skim — rebuilt with the point and its stakes in the first…

Run

/pm-os:skimmable-writing

Use when a long, meandering Slack thread needs to become a clear summary — the decisions, the action items with owners, and what's still open — read for *signal* against the reader's role and why they're catching up. Not for an IDEAS-format meeting summary (`meeting-summary`), reconstructing a work conversation's flow (`transcript-insights`), or pulling structured requirements from a discussion (`requirements-from-talk`).

Category

Communication, Writing & Influence

Output

Use when a long, meandering Slack thread needs to become a clear summary — the decisions, the action items…

Run

/pm-os:slack-signal

Use when practicing problem-analysis judgment by generating solutions that sound pragmatic but would actually fail — customer notes in, three plausible-but-flawed solutions out, each with the subtle reason it doesn't hold up. Not for triaging a real reported bug against committed work (`bug-triage`), or classifying a real decision's reversibility (`two-way-door`).

Category

Decisions & Prioritization

Output

Use when practicing problem-analysis judgment by generating solutions that sound pragmatic but would actually…

Run

/pm-os:solution-flaws

Use when requirements and a data-tables overview need turning into an optimized, readable SQL query with inline explanation of every major part. Not for defining the analytics events and schema to instrument in the first place (`tracking-schema`), or defining what metric to measure (`success-metric`).

Category

Metrics & Measurement

Output

Use when requirements and a data-tables overview need turning into an optimized, readable SQL query with…

Run

/pm-os:sql-from-spec

Use when a coding task needs one clean *mode* held without mixing — Staff Engineer mode for planning and architecture (no code), or Intern mode for precise, scoped implementation (no re-architecting). Not for producing a technical architecture brief from a PRD (`tech-arch-brief`), finding the application's shape or archetype (`eng-shape`), or translating a technical plan for business stakeholders (`tech-to-business`).

Category

Code/Engineering

Output

Use when a coding task needs one clean *mode* held without mixing — Staff Engineer mode for planning and…

Run

/pm-os:staff-or-intern

Navigate stakeholder complexity end-to-end — from power mapping and risk review through message framing, challenging meeting prep, difficult conversations, and executive presence review. Use when mapping power dynamics, planning influence strategy, or navigating organizational politics.

Category

Stakeholders & Politics

Output

Navigate stakeholder complexity end-to-end — from power mapping and risk review through message framing,…

Run

/pm-os:stakeholder

Use when you need to analyze and strategize on the power, interest, and influence of stakeholders in an initiative — an initiative and its stakeholder list in, a Power-Interest grid with an engagement plan out. Also `/stakeholder` Step 2, taking the power dynamics map from Step 1's `power-map`. Not for ranking raw formal/informal influence before this grid exists (`power-map`), auditing a specific PRD against this map for circulation risk (`stakeholder-risk`), or mapping motivation instead of power/interest (`hidden-agendas`).

Category

Stakeholders & Politics

Output

Use when you need to analyze and strategize on the power, interest, and influence of stakeholders in an…

Run

/pm-os:stakeholder-map

Use when you need to review a product requirement document or feature description for potential stakeholder risks and politics — a PRD or feature description in, political risks with mitigations and a readiness checklist out. Also `/stakeholder` Step 3, taking the Step 2 `stakeholder-map` snapshot as its starting point. Not for building the initial Power-Interest map from scratch (`stakeholder-map`), scanning for cross-functional objections before any specific doc exists (`objection-scan`), or ranking raw formal/informal influence (`power-map`).

Category

Stakeholders & Politics

Output

Use when you need to review a product requirement document or feature description for potential stakeholder…

Run

/pm-os:stakeholder-risk

Use when a vague work experience needs to become a tellable interview story — a raw achievement plus role, team and timeframe in, a polished Situation-Task-Action-Result story out, quantified, with your agency clear. Called by `pm-interview` Step 3 to build the stories its behavioral questions need. Not for anticipating the questions themselves (`pm-interview`), the résumé bullets (`resume`), or a written portfolio artifact (`work-sample`).

Category

Coaching, Self-Management & Career

Output

Use when a vague work experience needs to become a tellable interview story — a raw achievement plus role,…

Run

/pm-os:star-stories

Use when a user's answers to discovery questions (goal, process, pain points, time or money spent) need turning into 10 software product concepts, each one worth $30/month to that user. Not for problem+mechanism+differentiation product concepts from a stated problem area (`product-brainstorm`), or open-topic idea breadth (`brainstorm-genius`).

Category

Ideation & Analytical Thinking

Output

Use when a user's answers to discovery questions (goal, process, pain points, time or money spent) need…

Run

/pm-os:startup-ideas

Use when you need to write a status update that gets read and acted on — one status line, wins/next/risks/asks, sized to the audience and *skimmable* so the first three lines carry the point. Not for critiquing an existing update (`exec-update-review`), building a presentation deck (`problem-deck`), or choosing the week's priorities (`weekly-top-3`).

Category

Communication, Writing & Influence

Output

Use when you need to write a status update that gets read and acted on — one status line,…

Run

/pm-os:status-update

Use when you need to plan a presentation for a specific audience and time slot before building any slides — decide the depth, tone, data-vs-narrative balance, and the SCQA spine the deck will hang on. Not for writing the narrative prose (`presentation-narrative`), structuring slides from a problem (`problem-deck`), or reading the room politically (`prep-the-room`).

Category

Communication, Writing & Influence

Output

Use when you need to plan a presentation for a specific audience and time slot before building any slides —…

Run

/pm-os:strategic-deck

Use when a product situation needs Rumelt's three-part *kernel* — an honest diagnosis of what is actually going on, a guiding policy that chooses an approach, and the coherent actions that follow — built on a context check that refuses to proceed on fantasy inputs. Not for the seven-part strategy document (`product-strategy`), backcasting from an idealized end state (`limit-strategy`), or pointing existing strengths at one industry trend (`industry-strategy`).

Category

Strategy & Vision

Output

Use when a product situation needs Rumelt's three-part *kernel* — an honest diagnosis of what is actually…

Run

/pm-os:strategy-kernel

Use when a draft strategy document needs sharpening before a review — strengths and gaps named, specific copy/structure improvements written in the draft's own voice, and the top 3 questions the room will actually ask. Not for stress-testing an investment case's assumptions (`exec-deck`), or planning a deck before any slides exist (`strategic-deck`).

Category

Strategy & Vision

Output

Use when a draft strategy document needs sharpening before a review — strengths and gaps named, specific…

Run

/pm-os:strategy-review-prep

Use when a messy problem needs a clear recommendation and you do NOT need a full issue tree — a top-down (deductive) and bottom-up (inductive) pass that *converge* on one answer, delivered answer-first. Also /decisions Step 5, taking the MECE option tree and causal map as input and returning a Pyramid-Principle recommendation. Not for building the issue tree itself (`mckinsey-issue-tree`), a MECE decomposition (`mece-tree`), root-cause diagnosis alone (`causal-tree`), or framing the problem before analysis (`problem-framing-canvas`).

Category

Decisions & Prioritization

Output

Use when a messy problem needs a clear recommendation and you do NOT need a full issue tree — a top-down…

Run

/pm-os:structure-problem

Use when a design or product goal is vague — "improve onboarding," "make search better" — and the PM needs one primary metric with a baseline, a target, and *guardrails* before work starts. Not for the full measurement workflow (`/measure`), event-schema design (`tracking-schema`), estimating a feature's impact (`impact-sizing`), or designing the experiment (`experiment-design`).

Category

Metrics & Measurement

Output

Use when a design or product goal is vague — "improve onboarding," "make search better" — and the PM needs…

Run

/pm-os:success-metric

Use when someone wants to discover their one distinctive strength — interview them on past praise, who seeks their help, and what comes easily to them, then name the single superpower the evidence supports. Not for reframing a specific worry (`mindset-shift`), or a full PM performance coaching session (`coaching`).

Category

Coaching, Self-Management & Career

Output

Use when someone wants to discover their one distinctive strength — interview them on past praise, who seeks…

Run

/pm-os:superpowers

Use when a pile of open-ended survey responses needs turning into *pain points* with honest counts, verbatim quotes, and concrete action items. Fires on "analyze these survey results" or an exported CSV of free-text answers. Not for mixed-source research (`research-synthesis`), support tickets (`tickets-to-improvements`), or NPS programs (`nps-to-cx-plan`).

Category

Discovery & Research

Output

Use when a pile of open-ended survey responses needs turning into *pain points* with honest counts, verbatim…

Run

/pm-os:survey-to-actions

Use when a finished SWOT is sitting there doing nothing and needs turning into moves — all six quadrant *pairings* worked, each producing 5–10 reusable strategy patterns rather than company-specific tactics, then unified into themes. Not for generating growth ideas from a portfolio map (`pioneer-migrator-settler`), planning from current state through the projected mess to an ideal (`scenario-plan`), or counter-positioning ideation (`disruption-what-ifs`).

Category

Strategy & Vision

Output

Use when a finished SWOT is sitting there doing nothing and needs turning into moves — all six quadrant…

Run

/pm-os:swot-moves

Use when you have an existing list of tasks and need it turned into a sequenced plan — ordered by *dependency*, with the missing tasks surfaced, contingencies for the risky ones, and go/no-go milestones. Not for planning from a brief with a critical path (`project-plan`), ordering an unstructured to-do dump (`task-sequence`), or turning feedback into a release plan (`v2-plan`).

Category

Planning & Execution

Output

Use when you have an existing list of tasks and need it turned into a sequenced plan — ordered by…

Run

/pm-os:task-list-to-plan

Use when a chaotic to-do list needs organizing — dump everything, group by affinity, unpack anything too vague to act on, then sequence by dependency. Not for a project brief needing a critical path (`project-plan`), an already-clean task list needing dependency ordering (`task-list-to-plan`), or sequencing a small set of strategic components into a self-reinforcing loop (`reinforcing-sequence`).

Category

Planning & Execution

Output

Use when a chaotic to-do list needs organizing — dump everything, group by affinity, unpack anything too…

Run

/pm-os:task-sequence

Use when a PM wants to know how their designers and engineers actually see them — answers to 8 behavioral questions in, two simulated peer survey responses plus the named gap between self-image and outside view out. `/coaching` mode 2. Not for debriefing one real event (`situation-retrospective`), a systematic audit across all clusters (`blind-spot-scan`), or live practice against a frustrated colleague (`adversarial-roleplay`).

Category

Coaching, Self-Management & Career

Output

Use when a PM wants to know how their designers and engineers actually see them — answers to 8 behavioral…

Run

/pm-os:team-perspective-reveal

Use when you have a PRD and need a Technical Architecture Brief — system context, component architecture, tech-stack rationale, and risk mitigation — where every architecture *trade-off* names the requirement that drives it and the condition that would reverse it. Not for finding the application's shape or archetype (`eng-shape`), translating a decided technical plan for stakeholders (`tech-to-business`), or writing the PRD itself (`prd-draft`).

Category

Code/Engineering

Output

Use when you have a PRD and need a Technical Architecture Brief — system context, component architecture,…

Run

/pm-os:tech-arch-brief

Use when a technology announcement needs reading for what it changes about *your* situation — a neutral summary, a context block on the subject it lands on, then tail-sampled generation across four frames (new outcomes, new affordances, relative value, secret levers), critiqued down to a top three moves. Not for the sampling mechanics themselves (`verbalized-sampling`), scanning a market for inflection points (`inflection-scan`), running disruption scenarios against your own product (`disruption-what-ifs`), or getting oriented inside an unfamiliar codebase (`vibe-code-leaf-finder`).

Category

Ideation & Analytical Thinking

Output

Use when a technology announcement needs reading for what it changes about *your* situation — a neutral…

Run

/pm-os:tech-sensemaking

Use when you need to translate a technical explanation for a non-technical stakeholder so they can make *the decision* in front of them — jargon stripped, an analogy where it helps, risks preserved, and their likely questions answered. Not for producing the technical architecture itself (`tech-arch-brief`), presenting a deck to leadership (`problem-deck`), or writing a status update (`status-update`).

Category

Communication, Writing & Influence

Output

Use when you need to translate a technical explanation for a non-technical stakeholder so they can make *the…

Run

/pm-os:tech-to-business

Use when a batch of support tickets needs turning into a prioritized product-improvement backlog — recurring trends named and ranked by frequency, severity, and effort. Not for open-ended survey text (`survey-to-actions`), the full post-launch feedback-to-roadmap sequencing across all sources (`v2-plan`), or setting up the ongoing feedback pipeline itself (`feedback-loop`).

Category

Discovery & Research

Output

Use when a batch of support tickets needs turning into a prioritized product-improvement backlog — recurring…

Run

/pm-os:tickets-to-improvements

Use when a stakeholder decision is stuck on a conflict and you want to dissolve it with Theory of Constraints — trace symptoms to a root cause (Effect-Cause-Effect), surface the conflict and the assumption holding it (the Evaporating Cloud), then break that assumption. Not for a general root-cause tree (`causal-tree`), a MECE decomposition (`mece-tree`), or converging a messy problem to a recommendation (`structure-problem`).

Category

Stakeholders & Politics

Output

Use when a stakeholder decision is stuck on a conflict and you want to dissolve it with Theory of Constraints…

Run

/pm-os:toc-thinking

Use when you need to distill a stretch of completed work — todo list, calendar, shipped items — into the top 5 wins stated as outcomes, for a brag doc, perf review input, or end-of-week reflection. Not for planning the coming week (`weekly-top-3`), the weekly report to others (`status-update`), or interview-ready narratives (`star-stories`).

Category

Coaching, Self-Management & Career

Output

Use when you need to distill a stretch of completed work — todo list, calendar, shipped items — into the top…

Run

/pm-os:top-5-wins

Use when you have a UI or feature and metrics you want to measure, and need the analytics *events* and their schemas to instrument — each event tied to a real measurement question, with the properties needed to actually compute the metric. Not for defining what to measure (`success-metric`), querying data you already collect (`sql-from-spec`), or the full measurement workflow (`/measure`).

Category

Metrics & Measurement

Output

Use when you have a UI or feature and metrics you want to measure, and need the analytics *events* and their…

Run

/pm-os:tracking-schema

Use when you need to communicate a feature request's trade-off visually — place it on a priority×effort *2×2* against real comparisons so stakeholders see why it's worth pursuing, deferring, or rejecting. Not for prioritizing a whole requirements list (`p0-p1-p2`), pricing one scope-change request against a commitment (`scope-defense`), or ranking discovery opportunities (`ost-prioritize`).

Category

Decisions & Prioritization

Output

Use when you need to communicate a feature request's trade-off visually — place it on a priority×effort *2×2*…

Run

/pm-os:tradeoff-matrix

Use when a raw interview transcript needs a light readability pass before analysis — a raw transcript in, a clean one out with filler and false starts stripped, speaker voice and meaning untouched. Also `/research` Step 1. Not for extracting structured notes from the cleaned transcript (`interview-notes`), JTBD analysis of a customer interview (`interview-insights`), or writing the interview guide that precedes the interview (`mom-test-guide`).

Category

Discovery & Research

Output

Use when a raw interview transcript needs a light readability pass before analysis — a raw transcript in, a…

Run

/pm-os:transcript-cleanup

Use when a conversation transcript — a long meeting, a sprawling Slack thread, a call recording — needs *reconstruction*: who said what, how the discussion branched, what got committed, and what's still dangling. Fires on "what happened in this conversation" or "pull the decisions and actions out of this." Not for an IDEAS-format meeting summary (`meeting-summary`), customer-interview analysis (`interview-insights`), or extracting requirements (`requirements-from-talk`).

Category

Discovery & Research

Output

Use when a conversation transcript — a long meeting, a sprawling Slack thread, a call recording — needs…

Run

/pm-os:transcript-insights

Use when a decision needs classified as reversible or permanent before deciding how much process it deserves — a situation and decision in, a one-way/two-way-door classification and a recommended decision process out. Also `/decisions` Step 2, taking the Step 1 root cause map as input. Not for auditing a past decision (`decision-audit`), journaling this decision once classified (`decision-journal`), or checking whether external pressure is overriding your gut on a personal call (`gut-check`).

Category

Decisions & Prioritization

Output

Use when a decision needs classified as reversible or permanent before deciding how much process it deserves…

Run

/pm-os:two-way-door

Use when you have user stories and need testable, measurable UI acceptance criteria covering every state, breakpoint, and edge case — vague criteria like "looks good" or "works well" replaced with specifics QA can validate without ambiguity. Also `/prd` Step 5, only when the PRD has UI scope, taking Step 4's stories as input. Not for sweeping a brief for edge cases before stories exist (`ux-edge-cases`), resolving one specific edge case (`edge-case-balance`), or writing the underlying story these criteria attach to (`user-stories`).

Category

PRD, Specs & Documentation

Output

Use when you have user stories and need testable, measurable UI acceptance criteria covering every state,…

Run

/pm-os:ui-acceptance-criteria

Use when you have an existing UI screenshot and need it transcribed into structured JSON — every element, its position, style, and text content — for dev handoff, a design-system audit, or feeding another tool. Not for extracting PM-level requirements from a design (`requirements-from-design`), or sketching a flow that doesn't exist yet (`wireframe-sketch`).

Category

Design & Prototyping

Output

Use when you have an existing UI screenshot and need it transcribed into structured JSON — every element, its…

Run

/pm-os:ui-to-json

Use when a problem needs a business-level view before product development starts — who has it, what they do instead today, how often, why they'd switch, and how long they'd deliberate. Reforge's use-case-map method, a market/positioning lens distinct from product-development specs. Not for actor-goal-scenario specs for engineering (`use-cases`), or a single narrative problem frame (`problem-framing-canvas`).

Category

PRD, Specs & Documentation

Output

Use when a problem needs a business-level view before product development starts — who has it, what they do…

Run

/pm-os:use-case-map

Use when research notes or a raw product idea need to become Cockburn-style use cases — actor, goal, trigger, main success scenario, extensions and a testable success criterion each — validated against CRISP before anything downstream depends on them. Also /prd Steps 1c and 3. Not for extracting requirements from a meeting transcript (`requirements-from-talk`), a business-level actor-and-goal map before product work (`use-case-map`), agile role-goal-benefit stories (`user-stories`), or Given/When/Then acceptance criteria (`gherkin-stories`).

Category

PRD, Specs & Documentation

Output

Use when research notes or a raw product idea need to become Cockburn-style use cases — actor, goal, trigger,…

Run

/pm-os:use-cases

Use when a persona is buzzwords — "Sarah, 34, tech-savvy marketer" — and the PM needs it upgraded into an actionable profile: real *tasks*, constraints, and scenarios, with every claim either grounded in data or flagged as assumed. Not for building a first persona from research (`proto-persona`) or emotional/behavioral mapping (`create-empathy-maps`).

Category

Discovery & Research

Output

Use when a persona is buzzwords — "Sarah, 34, tech-savvy marketer" — and the PM needs it upgraded into an…

Run

/pm-os:user-profile

Use when you have a use-case list (or initiative requirements, if no use cases exist yet) and need agile-style role-goal-benefit user stories broken out and traceable back to their source. Also `/prd` Step 4a — the basic mode, alternative to Step 4b's Gherkin-embedded stories — taking Step 3's use-case list as input. Not for Gherkin/Given-When-Then acceptance criteria embedded in the story itself (`gherkin-stories`), translating stories into testable UI QA specs (`ui-acceptance-criteria`), or defining the use cases themselves (`use-cases`).

Category

PRD, Specs & Documentation

Output

Use when you have a use-case list (or initiative requirements, if no use cases exist yet) and need…

Run

/pm-os:user-stories

Use when a product brief needs a comprehensive edge-case sweep before build — a brief in, a scenario/expected-behavior table out, covering user types, contexts, failure inputs, load, integrations, security, and accessibility. Not for resolving one specific edge case already reported (`edge-case-balance`), or turning user stories into testable QA acceptance criteria (`ui-acceptance-criteria`).

Category

Design & Prototyping

Output

Use when a product brief needs a comprehensive edge-case sweep before build — a brief in, a…

Run

/pm-os:ux-edge-cases

Use when a written product requirement uses vague or incorrect UX terminology — a requirements doc or excerpt in, each imprecise term flagged with the correct UX/interaction term and why it matters, out. Not for evaluating a visual design (`product-design-analyzer`), or generating new design ideas (`affordances-signifiers`).

Category

Design & Prototyping

Output

Use when a written product requirement uses vague or incorrect UX terminology — a requirements doc or excerpt…

Run

/pm-os:ux-terminology

Use when a product has launched, feedback has piled up, and the PM needs a prioritized v2 plan — *themes* scored with one framework and sequenced into phases. Fires on "what should v2 be?" or "turn this feedback into a roadmap." Not for planning how to collect post-launch feedback (`feedback-loop`), answering research questions (`research-synthesis`), or support tickets alone (`tickets-to-improvements`).

Category

Discovery & Research

Output

Use when a product has launched, feedback has piled up, and the PM needs a prioritized v2 plan — *themes*…

Run

/pm-os:v2-plan

Use when a strategy needs checked against where value actually gets generated — a product/industry (or a strategy draft) in, a value chain from end-user needs to core value generators out, with the strategy's pillars checked against the leverage points that chain reveals. Also `/strategy` Step 3. Not for backcasting from an idealized end state (`limit-strategy`), scanning external trends for new opportunities (`inflection-scan`), or building the strategy document itself (`product-strategy`).

Category

Strategy & Vision

Output

Use when a strategy needs checked against where value actually gets generated — a product/industry (or a…

Run

/pm-os:value-chain

Use when the question is measure-more versus decide-now and the research spend needs a number — the *ceiling* any information could be worth, the value of one specific proposed measurement, and a MEASURE / DON'T MEASURE verdict against its cost. Also /measure Step 4, ranking which of the decomposed uncertainties to reduce first. Not for producing the estimate itself (`fermi-decomposition`), collecting the sample once measuring is justified (`rule-of-five`), or classifying how reversible the decision is (`two-way-door`).

Category

Metrics & Measurement

Output

Use when the question is measure-more versus decide-now and the research spend needs a number — the *ceiling*…

Run

/pm-os:value-of-information

Use when an output set needs genuine diversity and "give me 10 ideas" keeps collapsing into one cluster — prompt for a probability distribution with tail sampling instead of a single response, load context first, then critique the results before presenting them. Also the generation mechanics behind `tech-sensemaking` Step 3. Not for a facilitated ideation session (`brainstorm-genius`), constraint-driven idea generation (`constrained-ideas`), or generating questions rather than answers (`good-question-brainstormer`).

Category

Ideation & Analytical Thinking

Output

Use when an output set needs genuine diversity and "give me 10 ideas" keeps collapsing into one cluster —…

Run

/pm-os:verbalized-sampling

Use when a product, offering, doc, or flow is bloated and you want to improve it by *subtraction* — name the one core purpose, then cut every element that doesn't serve it. Not for cutting steps from a specific multi-step flow (`friction-reduce`), deciding what makes a release (`p0-p1-p2`), or pricing a proposed scope addition (`scope-defense`).

Category

Ideation & Analytical Thinking

Output

Use when a product, offering, doc, or flow is bloated and you want to improve it by *subtraction* — name the…

Run

/pm-os:via-negativa

Use when someone wants to know where in a codebase AI can write freely — a repo, package or directory in, every file classified LEAF (vibe it), BRANCH (guardrails) or TRUNK (human writes it) out, with stress tests per leaf. Fires on "vibe-code audit", "leaf finder", "where can I let Claude loose", "which files are safe to rewrite". Not for shaping an engineering approach with the team (`eng-shape`), writing up a technical architecture decision (`tech-arch-brief`), or getting oriented in an unfamiliar technical domain (`tech-sensemaking`).

Category

Code/Engineering

Output

Use when someone wants to know where in a codebase AI can write freely — a repo, package or directory in,…

Run

/pm-os:vibe-code-leaf-finder

Use when a vision or mission statement needs converting into a formal quarterly OKR set — objectives that are ambitious and qualitative, key results that are quantitative and tied to a real product milestone. Not for bridging a vision into a this-quarter execution plan with design principles (`vision-to-quarter`), or the underlying strategy the vision serves (`strategy-kernel`).

Category

Strategy & Vision

Output

Use when a vision or mission statement needs converting into a formal quarterly OKR set — objectives that are…

Run

/pm-os:vision-to-okrs

Use when a PM or design lead holds a vision deck and needs it *bridged* to this quarter's concrete work — themes turned into operational design principles, elements sorted by what's actionable now, and vague vision words pinned down. Not for writing OKRs from a vision (`vision-to-okrs`), a roadmap from user interviews (`now-next-later`), or the strategy itself (`strategy-kernel`).

Category

Decisions & Prioritization

Output

Use when a PM or design lead holds a vision deck and needs it *bridged* to this quarter's concrete work —…

Run

/pm-os:vision-to-quarter

Use when the week ahead is a pile of meetings, threads, and half-started projects and you need to triage it into the 3 priorities that define success — plus protected time to do them and a decision on what deliberately doesn't happen. Not for reviewing what you accomplished (`top-5-wins`), writing the weekly report (`status-update`), or roadmap horizons (`now-next-later`).

Category

Planning & Execution

Output

Use when the week ahead is a pile of meetings, threads, and half-started projects and you need to triage it…

Run

/pm-os:weekly-top-3

Use when you already have the content — findings, data, notes — and need to arrange it into a presentable flow, surfacing which of What / So What / Now What is missing. Not for deriving a storyline from a problem (`problem-deck`), writing narrative prose (`presentation-narrative`), or planning depth and tone for an audience (`strategic-deck`).

Category

Communication, Writing & Influence

Output

Use when you already have the content — findings, data, notes — and need to arrange it into a presentable…

Run

/pm-os:what-so-what-deck

Use when a product idea needs full flow sketches — multiple screens, the user journey between them, and friction points named at each touchpoint. Feeds `clickable-prototype` once a screen needs to be testable. Not for the single-screen kickoff sketch (`app-design-foundation`), a testable HTML prototype (`clickable-prototype`), or translating PM context into design constraints (`pm-to-design`).

Category

Design & Prototyping

Output

Use when a product idea needs full flow sketches — multiple screens, the user journey between them, and…

Run

/pm-os:wireframe-sketch

Use when one assumption needs the cheapest evidence that could settle it before anything gets built — a "We believe that…" statement in, five candidate signals and one chosen signal out with a method, a timebox, and a quantified pass threshold. Also `/assumptions` Step 3, run once per critical assumption. Not for generating the assumption set (`product-assumptions`), ranking which to test first (`risky-assumptions`), or designing a powered experiment once a signal justifies one (`experiment-design`).

Category

Discovery & Research

Output

Use when one assumption needs the cheapest evidence that could settle it before anything gets built — a "We…

Run

/pm-os:work-backwards

Use when you're building a PM work sample — a take-home, portfolio piece, or interview presentation — that has to show how you think and be *specific* to the company you're applying to. Not for writing the résumé (`resume`), preparing interview answers (`pm-interview`), building behavioral stories (`star-stories`), or drafting a real PRD (`prd-draft`).

Category

Coaching, Self-Management & Career

Output

Use when you're building a PM work sample — a take-home, portfolio piece, or interview presentation — that…

Run

/pm-os:work-sample

Use when messy evidence — interviews, transcripts, tickets, notes, a domain description — has to become a *representation* of how work actually functions before anyone proposes automation: the durable obligation forcing the work, its entities and states, where the truth is fragmented, and only then where AI can be inserted safely. Also fires on vertical SaaS exploration and "where does AI fit in this workflow". Not for mapping a system's UI and code affordances (`breadboarding`), the end-to-end customer experience across stages (`journey-map`), or generating product concepts from an opportunity (`product-brainstorm`).

Category

Discovery & Research

Output

Use when messy evidence — interviews, transcripts, tickets, notes, a domain description — has to become a…

Run

/pm-os:workflow-trellis

Use when a workshop has to be designed against hard constraints — a problem or a source transcript, a participant count, a time limit, and a platform in, either 3–5 activity variants to choose between or one sequenced agenda that reaches a named decision, with a minute-by-minute timeline that sums exactly to the time available. Not for a general pre-meeting brief (`meeting-outcomes`), a product refinement session's agenda (`refinement-session`), or running the ideation itself rather than designing the container for it (`brainstorm-genius`).

Category

Meetings & Facilitation

Output

Use when a workshop has to be designed against hard constraints — a problem or a source transcript, a…

Run

/pm-os:workshop-design