ReddRadar
Agent skill
diagnose
Validate hook integrity, state file consistency, and system health for the ~/.claude configuration.
Install this agent skill to your Project
npx add-skill https://github.com/juanandresgs/claude-ctrl/tree/main/skills/diagnose
SKILL.md
/diagnose — Hook & State Health Check
Validates the entire hook/state system is healthy. Run this when hooks behave unexpectedly, after configuration changes, or as a periodic sanity check.
Process
Step 1: Run the diagnostic script
bash ~/.claude/skills/diagnose/scripts/diagnose.sh
The script checks:
- Hook file integrity — all scripts referenced in settings.json exist and are executable
- Shared library health — log.sh and context-lib.sh source without errors
- State file validation — .plan-drift, .proof-status, .test-status formats
- Settings consistency — valid JSON, no duplicate hook registrations
- MASTER_PLAN.md status — phase statuses, DEC-ID format, REQ-ID format
- Git health — orphaned worktrees, uncommitted changes
Step 2: Interpret results
- PASS — check passed, no action needed
- WARN — non-critical issue; explain the risk to the user
- FAIL — critical issue; suggest specific remediation steps
Step 3: Remediation
For any FAILs, provide the user with:
- What is broken and why it matters
- The exact command or edit to fix it
- Whether the fix requires a worktree (Sacred Practice #2)
For WARNs, explain:
- What the risk is if left unaddressed
- Whether it is safe to ignore for now
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
juanandresgs/claude-ctrl
rewind
List and restore to a named checkpoint from the current session. Use when the agent has gone off track and you need to recover to a previous good state.
juanandresgs/claude-ctrl
consume-content
Produce a faithful content-snapshot of any source material (article, report, PDF, advisory) with verbatim quotes, structural transparency, and labeled editorial.
juanandresgs/claude-ctrl
deep-research
Multi-model deep research with comparative assessment (OpenAI + Perplexity + Gemini). Queries 3 deep research providers in parallel and produces a comparative synthesis.
juanandresgs/claude-ctrl
context-preservation
Generate structured context summaries for session continuity across compaction
juanandresgs/claude-ctrl
reckoning
Analyze a project's MASTER_PLAN.md to assess coherence, evolution trajectory, and intent alignment. Modes: default (full analysis), compare (delta between reckonings), operationalize (convert findings to actionable work via /decide), steer (strategic brainstorming grounded in findings).
juanandresgs/claude-ctrl
decide
Generate an interactive decision configurator from research or plan analysis. Presents options as explorable cards with trade-offs, costs, and filtering. Integrates with Planner to collect DEC-ID decisions.
Didn't find tool you were looking for?