PM-operating-OS
AGENTS.md3 starsChanged 18 days agoMITView on GitHub
PM-OS supports the same high-level workflows across both runtimes, but the packaging differs: These are available before you finish setup. These are generated by setup and personalized from config/pm-os-config.yaml. continual-learning is not yet a first-class Codex workflow. It remains Cursor-first / experimental because its current transcript-ingestion logic depends on Cursor transcript paths.
In this repository24 other files it gives its agents.
- AGENTS.mdAGENTS.md
- company-product-signalsSkill
- company-researcherSkill
- company-strategySkill
- pm-os-onboardingSkill
- brainstormingSkill
- continual-learningSkill
- decision-loggerSkill
- eval-briefSkill
- exec-communicatorSkill
- experiment-designerSkill
- experiment-writeupSkill
- knowledge-updaterSkill
- launch-postSkill
- launch-readinessSkill
- meeting-to-actionsSkill
- one-pagerSkill
- pptx-creatorSkill
- prd-writerSkill
- stakeholder-updateSkill
- strategy-connectorSkill
- what-ifSkill
- working-backwardsSkill
- writing-clearlySkill
The file
Files
PM-OS Workflows
PM-OS supports the same high-level workflows across both runtimes, but the packaging differs:
- Cursor uses generated agents under
~/.cursor/agents/ - Codex uses repo or user skills under
.agents/skills/and~/.agents/skills/
Bootstrap workflows
These are available before you finish setup.
| Workflow | Cursor surface | Codex surface | What it does |
| Onboarding | .cursor/agents/onboarding.md | .agents/skills/pm-os-onboarding/ | Collects PM context, writes config, runs setup |
| Company Researcher | .cursor/agents/company-researcher.md | .agents/skills/company-researcher/ | Merges company strategy and product-signal research |
| Company Strategy | .cursor/agents/company-strategy.md | .agents/skills/company-strategy/ | Researches company overview and strategy |
| Company Product Signals | .cursor/agents/company-product-signals.md | .agents/skills/company-product-signals/ | Infers product areas and goals from public signals |
Generated workflows
These are generated by setup and personalized from config/pm-os-config.yaml.
Feedback Analyzer
- Trigger examples: "analyze feedback", "customer feedback", "Slack feedback analysis"
- Requires: Slack MCP
- Behavior: analyzes feedback themes, compares to PM-OS memory, writes to
memory/feedback/
Weekly Planner
- Trigger examples: "plan my week", "daily focus", "plan my day"
- Requires: Slack MCP + Google Drive MCP
- Behavior: reads planning context, creates a prioritized plan, writes to
memory/weekly-plans/
Eval Brief
- Trigger examples: "write evals", "golden set", "quality bar", "how do we know this is good enough"
- Requires: a product dossier (or create
evals.mdfrom the template) - Behavior: turns AI quality into jobs, dimensions, judges, and a pass bar; writes
knowledge/products/<product>/evals.md
Strategy Reviewer
- Trigger examples: "strategy review", "check strategic fit", "does this align with our goals?"
- Requires: populated
knowledge/ - Behavior: scores PM artifacts against goals, pillars, and product principles
Exec Update Generator
- Trigger examples: "exec update", "status update", "SLT summary"
- Requires: best results with Slack + Google Drive MCP
- Behavior: compiles a BLUF executive update from PM-OS memory and live context
Retrospective
- Trigger examples: "retrospective", "what patterns do you see?", "strategy drift check"
- Requires: enough entries in
memory/ - Behavior: reads across memory and surfaces patterns, drift, blind spots, and recommendations
Runtime notes
Cursor
- Setup deploys generated agents to
~/.cursor/agents/ - Workflow invocation happens through Cursor chat and Cursor agent delegation
Codex
- Setup generates or syncs workflow skills into
.agents/skills/ - Workflow invocation happens through Codex skills, either implicitly or by explicit
$skill-name - Setup also generates
PM-OS-AGENTS.mdand.codex/config.toml
Codex v1 limitation
continual-learning is not yet a first-class Codex workflow. It remains Cursor-first / experimental because its current transcript-ingestion logic depends on Cursor transcript paths.
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.