product-manager-OS
CLAUDE.md4 starsChanged 2 months agoMITView on GitHub
Read AGENTS.md first. It is the single source of truth for how any agent runtime, this one included, operates this repository: load order, directory map, gate rules, and tool expectations. This file adds nothing beyond the router table below.
In this repository37 other files it gives its agents.
- AGENTS.mdAGENTS.md
- ai-prdSkill
- competitive-intelSkill
- conductorSkill
- decision-memoSkill
- design-reviewSkill
- escalationSkill
- experiment-designerSkill
- feedback-synthesisSkill
- gtm-launch-plannerSkill
- launch-readinessSkill
- market-sizingSkill
- metrics-treeSkill
- okr-criticSkill
- persona-builderSkill
- pm-hiringSkill
- postmortem-facilitatorSkill
- pricing-packagingSkill
- product-analystSkill
- product-reviewSkill
- program-premortemSkill
- reg-gap-checkSkill
- roadmap-builderSkill
- evidence-researchSkill
- experiment-analysisSkill
- independent-reviewSkill
- lifecycle-conductorSkill
- portfolio-planningSkill
- product-definitionSkill
- release-governanceSkill
- spec-reviewSkill
- stakeholder-updateSkill
- story-writerSkill
- strategy-criticSkill
- user-interviewSkill
- write-prdSkill
- write-vision-strategySkill
The file
We split this CLAUDE.md into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 3
Read AGENTS.md first. It is the single source of truth for how any agent runtime, this one included, operates this repository: load order, …
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md
CLAUDE
Read AGENTS.md first. It is the single source of truth for how any agent runtime, this one included, operates this repository: load order, directory map, gate rules, and tool expectations. This file adds nothing beyond the router table below.
Router
| When the user asks for | Invoke | Backing templates |
| "start", "start a product", or any wish to be interviewed through the loop | skills/conductor/SKILL.md; the full protocol is os/CONDUCTOR.md | templates/execution/state.md, the question banks in skills/conductor/questions/, and the stage templates each accepted answer lands in |
| "resume", "where are we", or "continue" on a product with a workspace | The resume protocol in os/CONDUCTOR.md: product README, then STATE.md, then the newest gate file, verify two answers, resume at the Next question ID | products/<name>/STATE.md |
| Evidence gathering for DISCOVER or metric evidence for OPERATE | skills/product-analyst/SKILL.md | templates/discovery/evidence-note.md, feeding the STATE.md evidence ledger |
| A PRD, a spec, a BRD or FRD, or "write up what we are building" | skills/write-prd/SKILL.md | templates/definition/prd.md, and the requirements stack around it: one-pager.md, brd.md, frd.md, nfr.md, business-rules.md, prfaq.md |
| A PRD for an AI-powered feature, or any spec where a model produces the output | skills/ai-prd/SKILL.md | templates/definition/prd.md plus the templates/ai/ overlay |
| "Is this spec any good", a requirements review, or a draft that reads clear but cannot be tested | skills/spec-review/SKILL.md | Any written spec; reports findings, never rewrites |
| A roadmap, quarterly plan, OKR set, or a stress-test of an existing one | skills/roadmap-builder/SKILL.md | templates/planning/roadmap.md, templates/planning/okrs.md |
| Raw feedback to work through: interview transcripts, support tickets, sales notes, reviews, survey text | skills/feedback-synthesis/SKILL.md | templates/discovery/user-research-plan.md, templates/discovery/discovery-document.md |
| Who this is for, a persona, a job story, a journey map, or an opportunity tree | skills/persona-builder/SKILL.md | templates/discovery/personas.md, jtbd-spec.md, journey-map.md, problem-framing.md, opportunity-solution-tree.md |
| Running a named method: "do a Kano", "score these with RICE", "run a premortem sheet", "map the job", "size the market", "SWOT this" | No skill needed for the sheet itself. Open frameworks/README.md, use the worksheet named there, and run its stated arithmetic unchanged | The worksheet's own Feeds list |
| "Is this a symptom or a structure", a failure that keeps coming back, or a metric that went flat while the team shipped steadily | No skill. Run frameworks/systems/iceberg-model.md to separate the symptom from what produces it, then frameworks/systems/causal-loop-diagram.md when the structure is a loop and frameworks/systems/leverage-points.md to place the fix on a rung. Diagnose before any planning sheet opens | templates/discovery/problem-framing.md, templates/execution/decision-log.md |
| "What kind of problem is this", or an argument about which method the work can even support | No skill. Run frameworks/systems/cynefin.md and let the domain pick the method: a clear-domain problem does not need a design sprint, and a chaotic one must not wait for a roadmap | templates/discovery/problem-framing.md, templates/execution/decision-log.md |
| "Are we set up to ship this", or whether the organization can carry the plan it just wrote | No skill. Run frameworks/assessment/product-operating-model-assessment.md, with frameworks/assessment/team-topologies-assessment.md for team shapes, frameworks/assessment/tech-debt-assessment.md for what the carried compromises cost, and frameworks/assessment/westrum-culture-typology.md for how bad news travels | templates/planning/capacity-plan.md, templates/planning/product-strategy.md |
| Planning or running customer interviews, and turning the notes into evidence | skills/user-interview/SKILL.md | templates/discovery/interview-guide.md, interview-notes.md, templates/discovery/evidence-note.md |
| A competitor teardown, a positioning question, or "how do we compare" | skills/competitive-intel/SKILL.md | templates/discovery/competitive-analysis.md, templates/planning/positioning.md |
| Market size, TAM, SAM, SOM, or "how big is this" | skills/market-sizing/SKILL.md | frameworks/strategy/market-sizing.md, templates/planning/business-case.md |
| Pricing, packaging, tiers, or willingness to pay | skills/pricing-packaging/SKILL.md | templates/planning/pricing-packaging.md, the frameworks/pricing/ sheets |
| A launch plan, go-to-market, or channel and comms sequencing | skills/gtm-launch-planner/SKILL.md | templates/planning/gtm-plan.md, templates/delivery/launch-comms-plan.md |
| An A/B test, an experiment, or "how would we know if this worked" | skills/experiment-designer/SKILL.md | templates/operate/experiment-brief.md |
| A north star, a metrics tree, a metric definition, or a dashboard | skills/metrics-tree/SKILL.md | templates/planning/north-star-metric.md, templates/operate/metrics-dictionary.md, dashboard-spec.md |
| An executive or board update, or a narrative that has to land in one read | skills/stakeholder-update/SKILL.md | templates/planning/exec-update.md, templates/operate/qbr-board-update.md |
| Epics, user stories, or acceptance criteria from a signed PRD | skills/story-writer/SKILL.md | templates/definition/acceptance-criteria.md, frameworks/prioritization/user-story-map.md |
| Drafting or critiquing OKRs | skills/okr-critic/SKILL.md | templates/planning/okrs.md |
| A vision, a product strategy, or "where are we going and what are we not doing" | skills/write-vision-strategy/SKILL.md | templates/planning/vision.md, templates/planning/product-strategy.md |
| A strategy to stress-test, or "is this actually a strategy" | skills/strategy-critic/SKILL.md | templates/planning/product-strategy.md, frameworks/strategy/strategy-kernel.md |
| A decision to write up, options to compare, or a one-way door to slow down | skills/decision-memo/SKILL.md | templates/planning/decision-memo.md, templates/execution/decision-log.md |
| An incident to review, a blameless postmortem, or a root cause to find | skills/postmortem-facilitator/SKILL.md | templates/operate/incident-postmortem.md, frameworks/execution/five-whys-fishbone.md |
| "Are we ready to ship", a go or no-go, or a Gate 5 walk | skills/launch-readiness/SKILL.md | templates/delivery/release-readiness.md, Gate 5 in os/STAGE-GATES.md |
| Hiring a PM: scorecard, interview loop, or calibration | skills/pm-hiring/SKILL.md | templates/execution/hiring-scorecard.md, knowledge/roles/pm-hiring-and-growth.md |
| "How much document does this need", or a request for a spec whose weight is unclear | Read os/WHICH-DOCUMENT.md and pick the weight, then skills/write-prd/SKILL.md drafts at that weight | templates/definition/one-pager.md or templates/definition/prd.md |
| A premortem, "what could kill this", or any risk pass before Gate 3 | skills/program-premortem/SKILL.md | templates/execution/risk-register.md, templates/execution/dependency-register.md |
| Anything touching a regulator, license condition, scheme rule, or compliance question | skills/reg-gap-check/SKILL.md | modules/regulated/ (byte-exact, never edited here) |
| A weekly product review, a WIP walk, or a review ritual that has decayed into status theater | skills/product-review/SKILL.md | templates/execution/decision-log.md, templates/operate/metrics-review.md |
| A stuck decision, a blocked dependency past its needed-by date, or a deadlock to escalate | skills/escalation/SKILL.md | templates/execution/risk-register.md, templates/execution/decision-log.md |
| "learn", "teach me", "quiz me", "study", or practice on a fictional product | learn/skills/tutor/SKILL.md; paths and workspace convention in learn/README.md | The path files in learn/, practice artifacts in learn/products/<name>/ |
| "Which domain", "domain pack", or what changes for a specific market | No skill. Read knowledge/domains/README.md and open the named card | The card's Templates-this-bends line |
"What does a <title> do", ladder, leveling, PM vs PMM, or role-scope questions | No skill. Read knowledge/roles/README.md and the card it routes to | templates/planning/first-90-days.md, templates/execution/stakeholder-map.md |
| A design review, a heuristic evaluation or cognitive walkthrough, a live-build design check before Gate 4, or an accessibility, content, localisation or deceptive-design pass on a flow | skills/design-review/SKILL.md | templates/architecture/design-review-record.md, templates/architecture/accessibility-checklist.md, templates/architecture/localisation-rtl-checklist.md, templates/architecture/design-md.md, templates/architecture/component-spec.md |
| A development handoff, a context package for engineering, or "is this ready to build" at the DESIGN exit | No skill. Read os/maps/design.md and fill templates/architecture/development-handoff.md, then run pmos handoff --path <product workspace> --product-id <id> --json to check the package and write handoff/context-index.json and handoff/CONTEXT.md | templates/architecture/development-handoff.md |
| Evidence gathering, drafting one template, validating a draft, attacking a draft, architecture options, acceptance testing, release management, metric analysis, growth diagnosis, positioning, or estimation | The matching instruction file in agents/ per the table in AGENTS.md; who leads which stage and what gets handed over is in agents/TEAM.md | Named per agent file |
"What does <term> mean here", or a word being used in a narrower sense than the industry uses it | No skill. Read GLOSSARY.md and follow the entry's link, since the governing file wins over the definition | The file the entry points at |
| "Why is this rule here", "isn't this waterfall", "why should I trust this", or how this compares to spec-kit, BMAD, a hosted PM product, or a template pack | No skill. Read docs/PHILOSOPHY.md for the belief and its counter-argument, docs/FAQ.md for the skeptical questions, docs/COMPARISON.md for the alternatives. Give the counter-argument too; a user who disagrees with a mechanism here is usually right about their own constraint | None. These files produce no artifact |
| Anything else in the product loop | No skill. Offer the Conductor once, per AGENTS.md load order step 0, then follow AGENTS.md: find the stage in os/OPERATING-LOOP.md, fill the stage's template from templates/, take it to its gate in os/STAGE-GATES.md | Per stage |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.