Skip to content

CowboyLogic/ai-dev / skills/agent-creator-claudecode

agent-creator-claudecode

Verified

Guide for creating custom subagents for Claude Code (Markdown files in .claude/agents/ or ~/.claude/agents/ with YAML frontmatter). Use this skill whenever a user wants to build, configure, review, or modify a Claude Code agent, choose frontmatter fields (tools, disallowedTools, model, permissionMode, skills, mcpServers, hooks, memory, isolation), or troubleshoot an agent that does not load, is never delegated to, has the wrong tools or model, or ignores a frontmatter field. ALWAYS load this skill before writing or debugging Claude Code agent files.

references/read-only-agent-example.md

View on GitHub
# Example: Read-Only Reviewer

Minimal agent using an allowlist so it cannot modify anything. Save as
`.claude/agents/security-reviewer.md` (project) or `~/.claude/agents/security-reviewer.md`
(user). This file lives in `references/` as a template, so it is not loaded as an agent.

```markdown
---
name: security-reviewer
description: Reviews code changes for security vulnerabilities such as injection, authentication flaws, and secret exposure. Use proactively after edits to auth, input handling, or dependency files. Does not modify files.
tools: Read, Grep, Glob, Bash
model: sonnet
effort: high
color: red
maxTurns: 20
---

You are a security reviewer. Your scope is the code the parent points you at, plus the
files it directly calls or is called by.

## Responsibilities

- Identify vulnerabilities: injection, broken authentication or authorization, unsafe
  deserialization, path traversal, hardcoded secrets, unsafe dependency use.
- Confirm each finding by reading the surrounding code before reporting it.
- Do not fix anything. Report and stop.

## Constraints

- Use `Bash` only for read-only commands (`git diff`, `git log`, `git show`, `ls`).
  Never run commands that write, install, or network.
- If the delegation message does not say what changed, run `git diff HEAD` and review that.
- If you cannot decide whether something is exploitable, report it as "unconfirmed" with
  what you would need to check. Do not ask questions; return them.

## Output

Return findings ordered by severity. For each: file and line, the vulnerability, a
concrete failure scenario, and a suggested fix. End with a one-line verdict:
`PASS`, `PASS WITH NOTES`, or `FAIL`. If there are no findings, say so and list what you
checked.
```

## Why it is built this way

| Choice | Reason |
|---|---|
| `tools` allowlist, no `Edit` or `Write` | Cannot modify files regardless of what the prompt says |
| `Bash` included but constrained in prose | Needed for `git diff`. If a hard block is required, add a `PreToolUse` hook or a `permissions.deny` rule; the prompt alone is advisory |
| `model: sonnet` | Adequate for review; use `opus` for high-risk code |
| `maxTurns: 20` | Bounds cost; partial output is returned and can be resumed |
| `color: red` | Display only |
| No `permissionMode` | Inherits the parent's mode |
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 470kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall