Skip to content

coroboros/agent-skills / skills/agent-creator

agent-creator

Verified

Create or update Claude Code subagent definitions in .claude/agents/, including prompts, tools, model inheritance and permissions. Use for agent configuration or Claude Code delegation mechanics, not routine delegation in another host.

references/debugging-agents.md

View on GitHub
# Debugging Claude Code agents

Start from the actual failing task and installed host. Use [official subagent documentation](https://code.claude.com/docs/en/sub-agents) for moving runtime details.

## Isolate the failure

Check whether the definition loads, the intended agent is selected, required tools are present, and the failure reproduces with the same input. Inspect scope precedence, malformed frontmatter, permissions and the selected model before changing the prompt. Preserve stderr and terminal status; accepted or queued work is not completion.

Reduce the failing input while retaining the behavior being investigated. Compare the final artifact with acceptance criteria rather than scoring the agent's explanation. For routing problems, include a near-miss request that should select another workflow.

## Observe only what is useful

Use host-provided task status, tool traces and error output. Avoid secrets in logs. Add temporary task-local diagnostics when they distinguish competing causes, then remove them when verified. Do not install dashboards, collectors or a retention system for a local debugging question.

A tool returning the same verdict on heterogeneous inputs needs its own negative control. Confirm the checker detects a known failure before trusting its clean result.

## Close the issue

Report the trigger, root cause supported by evidence, correction, verification and remaining limits. Preserve prior user changes. A successful prompt example is evidence for that case, not proof of general reliability.
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: 471kUpdated 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