Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/danielithomas/chealth/claude-mdgit clone --depth 1 https://github.com/danielithomas/chealthWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00756 | $0.00756 |
| Opus 5 | $0.00378 | $0.00378 |
| Sonnet 5 | $0.00151 | $0.00151 |
| Haiku 4.5 | $0.00076 | $0.00076 |
Grade A, and why
chealth CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Project Purpose
chealth is a CLI tool and Claude Code plugin that checks CLAUDE.md files against Anthropic best practices (20 rules). The CLI runs deterministic checks; the plugin skill adds semantic analysis powered by Claude.
Language & Tooling
- Language: Python 3.10+
- Package manager: uv
- Build backend: hatchling
Build & Test
uv run chealth .— check this project's CLAUDE.mduv run chealth --format json .— JSON output for automationuv run chealth --check line-count .— run a specific check onlyuv run pytest— run all testsuv run pytest tests/test_checks/test_line_count.py— run a single test file
Architecture
Check System
Checks implement a Check Protocol (structural typing, no inheritance) with name, description, rule_reference attributes and a run(ctx: CheckContext) -> list[CheckResult] method. Each check lives in its own module under src/chealth/checks/ and self-registers via @register() at import time. All check modules must be imported in cli.py:_load_checks() to trigger registration.
Two-Phase Execution (runner.py)
- Priority checks always run first regardless of
--checkfilters:file-discoverypopulatesctx.discovered_files,import-resolutionpopulatesctx.import_targets. These hydrateCheckContextfor all subsequent checks. - Standard checks run after context is populated, respecting any
--checkfilter.
Exit codes: 0 = all pass, 1 = worst is WARN, 2 = worst is DANGER. Severity is an IntEnum enabling max() to compute the worst result.
Key Utilities (checks/__init__.py)
read_file_safe(path)— tries utf-8, utf-8-sig, latin-1 encodingsstrip_code_blocks(content)— removes fenced code blocks to prevent false positives in pattern-matching checks
Plugin Integration
.claude-plugin/plugin.json declares the skill. skills/chealth/SKILL.md defines the plugin workflow: run deterministic CLI checks (JSON), then perform semantic analysis using skills/chealth/reference.md (the 20-rule guide). Each check's rule_reference traces back to a specific best-practice rule.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 68 lines · 756 tokens per session scan A 21e9370355b9
chealth CLAUDE.md is an instructions file published in the GitHub repository danielithomas/chealth (2 stars, last pushed 5mo ago), licensed MIT. It adds 756 tokens to every session, about $0.0038 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
claude-code_rails-upgrade-skill CLAUDE.md
Instructions for ombulabs/claude-code_rails-upgrade-skill, covering project conventions for claude, repository tooling, version guides (rails-upgrade/version-guides/.md), assigning priority (🔴 high / 🟡 medium / 🟢 low) and 🔴 high.
recursive-decomposition-skill AGENTS.md
Instructions for massimodeluisa/recursive-decomposition-skill, covering agents: recursive-decomposition skill, read these first (mandatory), language policy, non-negotiables and commands.
csift AGENTS.md
Instructions for wdhwg001/csift, covering agents.md - csift operating manual, 1. what csift is, 3.1 data location + path encoding, 3.2 record model (one json object per line) and 3.4 askuserquestion.
self-review CLAUDE.md
Claude Code instructions for HoussemDjeghri/self-review, covering self-review — working notes for this repository, plugin/ is the product; everything beside it is not, ./test.sh is the gate, claims in the docs are enforced, not asserted and releasing.
veriflow-cc CLAUDE.md
Claude Code instructions for bjwanneng/veriflow-cc, covering claude.md, project: veriflow-cc, mandatory: plan before code, mandatory: tdd (test-driven development) and test location.
timbogp-marketplace CLAUDE.md
Instructions for TimboGP/timbogp-marketplace: This file is kept as a compatibility pointer for Claude Code.