agentleFS
jMertajMerta/codex-skills/manage-product-ideas/SKILL.md

manage-product-ideas

Skill130 starsChanged 11 months agoMITView on GitHub

Use when the user asks to capture, clarify, document, audit, generate, compare, prioritize, validate, or realize product ideas. Guides one-question-at-a-time discovery, evidence-based scoring, explicit readiness decisions, and scoped delivery without assuming a specific product, repository layout, or planning plugin.

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    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
  1. Distinctive:

    Read references/idea-workflow.md before using any routed mode.

    line 10
  2. Distinctive:

    Ask exactly one unresolved product decision.

    line 30
  3. Distinctive:

    Wait for the user's answer.

    line 32
  4. Distinctive:

    Update the idea document immediately when documentation changes were requested.

    line 33
  5. Distinctive:

    Do not ask the user for facts that available sources can answer.

    line 36
  6. Distinctive:

    Prefer a bounded first useful outcome over platform work without a concrete customer case.

    line 44
  7. Distinctive:

    Do not treat a numerical score as stronger than its evidence.

    line 47

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill 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
SKILL.md

Manage Product Ideas

Turn a fuzzy opportunity into a documented, evidence-backed decision. Treat current code, tests, contracts, analytics, and operations as truth about the product; treat idea documents and Git history as decision context.

Read references/idea-workflow.md before using any routed mode.

Route the request

Choose the narrowest applicable mode:

RequestMode
Raw or fuzzy ideaGRILL
Create or revise an idea documentDOCUMENT
Review an idea portfolioAUDIT
Propose new directionsGENERATE
Choose what to do nextSELECT
Build an approved ideaREALIZE

For combined requests, use only applicable modes in this order: AUDIT → GENERATE → GRILL → DOCUMENT → SELECT → REALIZE.

Grill before judging

  1. Inspect available repository, product, customer, and operational evidence.
  2. Ask exactly one unresolved product decision.
  3. Include a recommended answer and its primary trade-off.
  4. Wait for the user's answer.
  5. Update the idea document immediately when documentation changes were requested.
  6. Continue until critical decisions are resolved and both sides share the same understanding.

Do not ask the user for facts that available sources can answer. Do not dump a questionnaire. Do not score, select, plan, or implement while critical branches remain unresolved.

Discover the repository's existing product-idea convention before creating files. If none exists and documentation is requested, default to docs/product-ideas/<slug>.md. Do not create a parallel hierarchy without a reason.

Preserve decision quality

  • Separate facts, assumptions, recommendations, confidence, and next actions.
  • Recover why a removed capability disappeared before proposing its return.
  • Prefer a bounded first useful outcome over platform work without a concrete customer case.
  • Below 70% confidence, choose a validation step instead of implementation.
  • Generate at most three candidates by default and persist only the candidates the user chooses.
  • Do not treat a numerical score as stronger than its evidence.

Gate realization

Before implementation, require:

  • a defined user and problem;
  • meaningful evidence or an explicit validation hypothesis;
  • a measurable outcome and stop condition;
  • a bounded first release and non-goals;
  • identified dependencies, trust boundaries, risks, rollout, and rollback;
  • resolved critical product rules.

Implementation, commits, pushes, and pull requests require their normal authorization. A selected idea or plan does not by itself authorize code changes.

Hand off

Return the current decision, confidence, strongest evidence, largest unknown, next action, and any document changed. When comparing candidates, explain why the runner-up loses instead of presenting scores without a decision.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.