agentleFS
garbageekgarbageek/claude-plugins/skills/init-repo-instructions/SKILL.md

init-repo-instructions

Skill1 starsChanged 35 days agoMITView on GitHub

Create repository-specific agent instructions for the requested Claude or Codex scope from real commands, boundaries, and non-obvious constraints. Preserve existing guidance; use when initializing instructions, not transforming the whole repository.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are 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.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read discovery and retention before choosing a target file.

    line 8
  2. Distinctive:

    Read existing instructions, manifests, commands, workspace boundaries, generated/manual sources, and relevant implementation evidence.

    line 15
  3. Distinctive:

    If the requested file is missing and creation is explicitly authorized, create that one file.

    line 23
  4. Distinctive:

    If useful instructions already exist, do not replace or rearrange them to impose a preferred skeleton.

    line 25
  5. Distinctive:

    Read back any authorized edit and inspect the diff.

    line 28
  6. Distinctive:

    Use review when the task is primarily to assess an existing file.

    line 32

≈ “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

Initialize Repository Instructions

Read discovery and retention before choosing a target file. It owns the shared evidence rubric and host-specific loading rules; do not invent a second precedence policy here.

  1. Resolve the requested host/scope and existing instruction chain. Prefer the established file and actual loading mechanism. Do not automatically generate both AGENTS.md and CLAUDE.md, a bridge, local overrides, or nested files.
  2. Read existing instructions, manifests, commands, workspace boundaries, generated/manual sources, and relevant implementation evidence. Identify mistakes an agent would otherwise make; omit generic advice and directory inventories. Use the repository's own project type and terminology.
  3. Draft only supported, actionable rules: actual commands with necessary working-directory/prerequisite details, responsibilities, unusual conventions, generated-file handling, and verified recurring pitfalls. No invented service, branch, deployment, test-creation, or documentation policy.
  4. If the requested file is missing and creation is explicitly authorized, create that one file. A request for a draft/proposal returns content without writes. If useful instructions already exist, do not replace or rearrange them to impose a preferred skeleton; propose the smallest update unless editing is already authorized. Do not re-ask permission for the exact approved edit.
  5. Read back any authorized edit and inspect the diff. Preserve valuable original rules and headings, omit empty/unverified sections, and report the actual file changed. Describe loading as unverified unless the host really confirmed it.

Use review when the task is primarily to assess an existing file. No ancillary reports, inventories, configuration changes, ready-made file matrix, hooks, or all-purpose “AI-ready” operation are part of initialization.

For repository-specific GitHub Copilot Code Review customization, use copilot-review-customizer. That workflow selects Copilot's review surfaces and uses review/regression history; it does not change this skill's Claude/Codex initialization or instruction-review contract.

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.