agentleFS
dnlboxdnlbox/ai-protocol/.agents/skills/shape-concept/SKILL.md

shape-concept

Skill18 starsChanged 3 months agoMITView on GitHub

Turn an accepted intake into concise canonical product intent without planning implementation.

The file

Show its rules and text

We split this skill into the rules it gives an agent

23 rules. About 0 are common, 18 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈18

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill after concept intake or when accepted product intent must change deliberately.

    line 8
  2. Very specific:

    Read every existing file in docs/concept/.

    line 13
  3. Distinctive:

    Treat placeholder files as scaffolding.

    line 16
  4. Distinctive:

    Treat accepted concept files as product authority.

    line 17
  5. Distinctive:

    Treat intake as historical input during re-shaping.

    line 18
  6. Very specific:

    Keep docs/concept/README.md as the concept index.

    line 25
  7. Very specific:

    Use status Shaping during initial drafting.

    line 26
  8. Very specific:

    Use status Re-shaping during deliberate revision.

    line 27
  9. Distinctive:

    Avoid a fixed document taxonomy.

    line 28
  10. Distinctive:

    Avoid one file per minor topic.

    line 29
  11. Distinctive:

    Avoid duplicating a decision across files.

    line 30
  12. Distinctive:

    Keep each discussion round focused on one tension.

    line 42
  13. Distinctive:

    Ask for one material product decision at a time.

    line 43
  14. Distinctive:

    Avoid detached research dumps.

    line 46
  15. Very specific:

    Write only within docs/concept/.

    line 50
  16. Distinctive:

    Update intake status only when the stage changes.

    line 51
  17. Distinctive:

    Avoid choosing frameworks during concept shaping.

    line 52
  18. Distinctive:

    Avoid choosing schemas during concept shaping.

    line 53
  19. Distinctive:

    Avoid choosing deployment topology during concept shaping.

    line 54
  20. Distinctive:

    Avoid turning open product questions into implementation discretion.

    line 55
  21. Distinctive:

    Stop before implementation planning.

    line 56
  22. Distinctive:

    Stop before product implementation.

    line 57
  23. Distinctive:

    Ask the user to accept the shaped concept.

    line 71

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Shape concept

Use this skill after concept intake or when accepted product intent must change deliberately.

Start

  • Read docs/intake/concept-intake.md.
  • Read every existing file in docs/concept/.
  • Require intake status Ready for concept shaping for initial shaping.
  • Require an explicit product-change request for re-shaping.
  • Treat placeholder files as scaffolding.
  • Treat accepted concept files as product authority.
  • Treat intake as historical input during re-shaping.
  • Return to concept intake when the intake is materially incomplete.

Document shape

  • Choose the smallest useful concept document set.
  • Give each concept file one ownership boundary.
  • Keep docs/concept/README.md as the concept index.
  • Use status Shaping during initial drafting.
  • Use status Re-shaping during deliberate revision.
  • Avoid a fixed document taxonomy.
  • Avoid one file per minor topic.
  • Avoid duplicating a decision across files.

Drafting

  • Synthesise the intake by meaning.
  • Separate settled intent from assumptions.
  • Separate evidence-backed findings from product judgements.
  • Separate explicit exclusions from open questions.
  • Cite factual claims with links or local sources.
  • Mark unsupported product judgements as hypotheses.
  • Preserve compatible human additions.
  • Preserve unresolved non-blocking questions.
  • Keep each discussion round focused on one tension.
  • Ask for one material product decision at a time.
  • Conduct research only when the user requests it.
  • Fold approved research into the relevant concept file.
  • Avoid detached research dumps.

Boundaries

  • Write only within docs/concept/.
  • Update intake status only when the stage changes.
  • Avoid choosing frameworks during concept shaping.
  • Avoid choosing schemas during concept shaping.
  • Avoid choosing deployment topology during concept shaping.
  • Avoid turning open product questions into implementation discretion.
  • Stop before implementation planning.
  • Stop before product implementation.

Completion

  • Require a coherent problem definition.
  • Require a coherent audience definition.
  • Require a coherent desired outcome.
  • Require coherent primary journeys.
  • Require explicit scope boundaries.
  • Require explicit non-negotiables.
  • Require explicit exclusions.
  • Require material domain rules.
  • Require material contradictions to be resolved.
  • Allow implementation-level unknowns to remain open.
  • Ask the user to accept the shaped concept.
  • Set concept status to Ready for operational planning after initial acceptance.
  • Set concept status to Ready for operational re-planning after re-shaping acceptance.
  • Set intake status to Concept shaped after initial acceptance.
  • Return a one-screen handoff summary.

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.