Chealth
Check CLAUDE.md health against Anthropic best practices — runs deterministic CLI checks then semantic analysis
Install
$ agentstack add skill-danielithomas-chealth-chealth ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-danielithomas-chealth-chealth) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
/chealth — CLAUDE.md Health Check
You are performing a comprehensive health check on this project's CLAUDE.md files. Follow these steps exactly.
Step 1: Run deterministic checks
Run the chealth CLI to get structured findings:
uv run --project "${CLAUDE_PLUGIN_ROOT}" chealth --exit-zero --format json . 2>&1If the CLI is not installed, tell the user: > chealth plugin not available. Reinstall with: claude plugin add chealth
Parse the JSON output. Note all WARN and DANGER findings.
Step 2: Read all discovered CLAUDE.md files
From the CLI JSON output, read every file listed in the file-discovery result's details.files array. Also read any .claude/rules/*.md files and files referenced by @import paths.
Step 3: Perform semantic checks
Using the best practices in ${CLAUDE_SKILL_DIR}/reference.md as your reference, evaluate:
- Contradiction detection — Identify rules that conflict with each other, including across files (Rule 11)
- Vague instruction assessment — Evaluate whether instructions are concrete enough to verify (Rule 9)
- Instruction effectiveness — Judge whether Claude would actually follow each instruction (Rule 18)
- Scope appropriateness — Determine if content belongs in root vs subdirectory vs
.claude/rules/(Rules 2, 15, 20)
Step 4: Present unified report
Combine deterministic and semantic findings into a single report using this format:
## CLAUDE.md Health Report
### Deterministic Checks (chealth CLI)
[For each finding:]
- [PASS|WARN|DANGER] **check-name**: message [Rule N]
### Semantic Analysis
[For each finding:]
- [PASS|WARN|DANGER] **check-name**: description
### Summary
- X checks passed, Y warnings, Z dangers
- [Top recommendation if any WARN/DANGER found]Use PASS/WARN/DANGER severity levels consistently. Semantic findings are advisory — label them as such.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: danielithomas
- Source: danielithomas/chealth
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.