| name | camel-debug |
| description | Ad-hoc troubleshooting for broken Camel routes outside of a pipeline run. |
| user_invocable | false |
Camel Debug — Standalone Troubleshooting
Diagnose and fix a broken Camel route using a structured STOP → PRESERVE → DIAGNOSE → FIX → GUARD workflow.
Announce: "Let me debug this using the camel-debug skill."
When to Use
- A route was working but is now broken (startup failures, runtime exceptions, unexpected behavior)
- A user arrives with "my route is broken" outside of an active pipeline run
- Ad-hoc troubleshooting — no pipeline context required
When NOT to Use
- Build/test failures during pipeline execution —
camel-verify handles this (dispatched by camel-execute)
- Quality validation of generated routes — use
/camel-validate
- Designing new integrations — use
/camel-brainstorm
- Planning or executing implementation — use
/camel-plan or /camel-execute
Violating the letter of these rules is violating the spirit of these rules.
Prerequisites
.camel-kit/config.properties must exist (for runtime and version detection)
- The project must have existing route files to debug
Guides
| Guide | When to Load | Purpose |
|---|
guides/debug-workflow.md | Always | Full debugging workflow — STOP, PRESERVE, DIAGNOSE, FIX, GUARD |
camel-verify/guides/error-taxonomy.md | Always (reference) | Error classification tables — patterns, categories, fix actions |
Shared Guides
Load these shared guides at workflow start:
shared/iron-laws.md — Iron Law 1 (MCP verification) applies during diagnosis
shared/mcp-setup.md — MCP tool configuration for catalog queries
Subagent Isolation
Diagnosis steps (Step 3 in the workflow) MUST run as subagents to keep diagnostic output out of the main context. Only the classification result and fix recommendation flow back to the main conversation.
| Subagent | Purpose | Input | Output |
|---|
| Route analyzer | Inspect route YAML for structural issues | Route file paths, error message | Structural findings list |
| MCP verifier | Verify components/endpoints against catalog | Component names from routes | Verification results (exists/missing/wrong options) |
| Log analyzer | Parse and classify error output | Raw log/stack trace text | Classified error with taxonomy match |