agentleFS
CsHengCsHeng/agent-skills/skills/analyze-project/SKILL.md

analyze-project

Skill3 starsChanged 8 days agoMITView on GitHub

Answer read-only project-state questions from current repository truth, from a local fact to requested broad orientation or a truth audit. Expand investigation when evidence is insufficient; not for implementation or code review. Supply bounded project evidence to a more specific domain diagnosis.

6 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When ownership is unclear, stable sources conflict, or evidence is insufficient, expand along that gap: inspect the relevant repository …

    line 13
  2. Distinctive:

    When another Skill owns the response (implementation, docs reorganization, change review, or a runtime, infrastructure, security, or other …

    line 16
  3. Distinctive:

    Read enough stable evidence to support the answer and verify targeted claims read-only.

    line 22
  4. Distinctive:

    When document health or drift affects confidence or is requested, read Document Health And Drift and classify the affected scope.

    line 23
  5. Very specific:

    Use project as the analysis unit, and keep terminology repository-local unless an applicable stable owner defines a wider convention.

    line 28
  6. Distinctive:

    Keep this Skill read-only.

    line 30

“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “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

Analyze Project

Answer the actual project question from sufficient current evidence, scaling both the investigation and the answer to the scope asked.

Select The Scope

  • For a local fact, identify the relevant project, follow applicable scoped instructions, read the stable owner that can answer the question, and verify the specific claim from code, configuration, or commands when needed.
  • When ownership is unclear, stable sources conflict, or evidence is insufficient, expand along that gap: inspect the relevant repository maps, README/AGENTS, ignore rules, or implementation, and state uncertainty rather than guessing. When documentation is poor, reconstruct from code for the asked question only.
  • For explicitly requested broad orientation or a full truth audit, map the relevant stable truth and search boundaries before drawing conclusions, and use Full Project Truth Audit.

When another Skill owns the response (implementation, docs reorganization, change review, or a runtime, infrastructure, security, or other domain diagnosis), contribute only project facts, confidence, and drift evidence.

Investigate And Answer

  1. Bound the question and the project. Honor the most specific applicable project instructions, and use docs/AGENTS.md, AGENTS.md, docs/README.md, and README.md to locate the truth owner.
  2. Separate stable truth from stage history before searching. Search stage history when requested or when a gap stable truth cannot answer needs it, and say so when used. Check ignore rules and generated or ignored paths before concluding something is absent.
  3. Read enough stable evidence to support the answer and verify targeted claims read-only. Escalate when contradictions or missing facts prevent a reliable conclusion.
  4. When document health or drift affects confidence or is requested, read Document Health And Drift and classify the affected scope.
  5. Give the conclusion, relevant evidence, and any material limitation once. Read Output Contract when reporting a broader truth map, drift, or a requested audit.

Boundaries

  • Use project as the analysis unit, and keep terminology repository-local unless an applicable stable owner defines a wider convention.
  • Preserve the subject and limits of each claim: file existence, documented intent, current implementation, observed runtime behavior, and inferred cause are different evidence.
  • Keep this Skill read-only; document reorganization or other mutation needs its own authorized request.

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.