agentleFS
Sign inStart free
axrmxvaxrmxv/claude-config/skills/add-agent/SKILL.md

add-agent

Skill1 starsChanged 4 months agoMITView on GitHub

Scaffold a new subagent for this config repo with least-privilege tools and register it in the authoritative roster. Use when adding a reviewer or specialist agent.

9 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use setup.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • VercelSkill

    setup

    Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Add a subagent named $ARGUMENTS to this repository.

    line 8
  2. Distinctive:

    Check it does not already exist: If an existing agent covers the job, extend that one instead — two agents with overlapping descriptions …

    line 18
  3. Very specific:

    Least privilege on tools: Read-only agents get Read, Grep, Glob, Bash and nothing else.

    line 37
  4. Very specific:

    Only grant Write/Edit to an agent whose job is to change files.

    line 38
  5. Very specific:

    Register it in the roster: Add a row to the table in .claude/rules/common/agents.md (Agent | Purpose | When to Use).

    line 45

“Common” is measured: kept by at least 6 in 10 of the 7 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

Add a subagent named $ARGUMENTS to this repository.

Existing agents

ls agents/ 2>/dev/null || ls .claude/agents/

Steps

  1. Check it does not already exist. If an existing agent covers the job, extend that one instead — two agents with overlapping descriptions make the delegation choice ambiguous.

  2. Create agents/<name>.md with this frontmatter:

    ---
name: <name>
description: <when to use it — this is what the orchestrator matches on>
model: <haiku | sonnet | opus>
tools: <comma-separated>
---
  3. Pick the model by cost of being wrong, per .claude/rules/common/efficiency.md:
    • haiku — light, mechanical work (documentation updates).
    • sonnet — the default for workers (review, tests, build fixes, refactors).
    • opus — deep reasoning where a miss is expensive (architecture, security).
  4. Least privilege on tools. Read-only agents get Read, Grep, Glob, Bash and nothing else. Only grant Write/Edit to an agent whose job is to change files. A reviewer that can edit code will quietly fix what it should have reported.
  5. Write the body as: what the agent is, its process, and what it outputs. Reviewers state severity levels and an explicit verdict. Keep it short — an agent prompt that rambles gets skimmed the same way a long CLAUDE.md does.
  6. Register it in the roster. Add a row to the table in .claude/rules/common/agents.md (Agent | Purpose | When to Use). That table is authoritative: the review rules forbid invoking an agent that is not listed there, so an unregistered agent is dead weight.
  7. Update the docs that list agents (docs/, docs/ru/, README) and run /audit-config to confirm the roster and agents/ are in sync.

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.

Check your setup, free