agentleFS
tenforward-audiotenforward-audio/swe-harness/.agents/skills/investigate-project/SKILL.md

investigate-project

Skill0 starsChanged 3 months agoMITView on GitHub

Answer, explain, diagnose, research, compare, plan, or audit this project without changing product behavior or work tracking. Use for root-cause analysis, architecture or implementation questions, feasibility and option comparisons, and requested plans. Do not use for a technical review of a defined candidate change, explicitly framed intake, implementation, queue changes, or release actions.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    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:

    Only the active agent responsible for the user-facing outcome applies this mode.

    line 12
  2. Distinctive:

    Read the root AGENTS.md, then only the files and policies needed for the question.

    line 19
  3. Distinctive:

    Do not edit files, dependencies, queues, status boards, branches, commits, releases, or external systems.

    line 33
  4. Distinctive:

    Do not treat repository history as proof that an old data or API shape needs support.

    line 35
  5. Very specific:

    If the request explicitly includes implementation, hand the evidence to develop-project for the authorised change.

    line 42
  6. Distinctive:

    If the user decides that an isolated planning or research idea is not worth pursuing, summarise the findings, uncertainty, and any material …

    line 48
  7. Distinctive:

    Keep the investigation read-only until the user unambiguously agrees to the named findings record, any lifecycle effect, and the local …

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Investigate the Project

Execution mode

delegate-readonly

Only the active agent responsible for the user-facing outcome applies this mode. A worker already delegated an investigation executes its assigned scope inline and must not dispatch another subagent. If the scope is too broad or blocked, return the evidence and a proposed split to the active agent.

Establish evidence read-only

  1. Read the root AGENTS.md, then only the files and policies needed for the question.
  2. State the question, current behavior, and evidence needed to answer it.
  3. Inspect code, tests, logs, history, configuration, and generated output with read-only commands. Run non-mutating checks when they materially reduce uncertainty.
  4. Trace the owning boundary and distinguish observed facts, supported inferences, assumptions, and unknowns.
  5. Report the answer, likely cause, risks, or options with concrete file and behavior evidence. For a plan, name outcomes, dependencies, verification, and decisions without creating tracking state.

Keep the boundary strict

  • Do not edit files, dependencies, queues, status boards, branches, commits, releases, or external systems.
  • Do not treat repository history as proof that an old data or API shape needs support; identify the concrete current compatibility obligation.
  • For audits, lead with actionable correctness, security, data-loss, compatibility, concurrency, failure-behavior, and test findings. Say plainly when no material issue is found and note residual uncertainty.
  • A review of a defined candidate commit, branch, worktree, or patch belongs to review-project-change.
  • If the request explicitly includes implementation, hand the evidence to develop-project for the authorised change. Explicitly framed support tickets and feature requests remain capture-only.

Offer deliberate abandonment without inferring it

  • If the user decides that an isolated planning or research idea is not worth pursuing, summarise the findings, uncertainty, and any material local work.
  • Name the retained worktree and branch and offer to preserve useful findings and clean up those exact local targets through clean-up-worktree. Research completion or a negative recommendation alone is not abandonment or cleanup authority.
  • Keep the investigation read-only until the user unambiguously agrees to the named findings record, any lifecycle effect, and the local cleanup.

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.