Borrowing it
Nothing to install: this file belongs to redhat-community-ai-tools/harness-eval. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/redhat-community-ai-tools/harness-eval/main/CLAUDE.mdgit clone --depth 1 https://github.com/redhat-community-ai-tools/harness-evalWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/redhat-community-ai-tools/harness-eval/claude-md)<a href="https://agentmods.dev/instructions/redhat-community-ai-tools/harness-eval/claude-md"><img src="https://agentmods.dev/badge/instructions/redhat-community-ai-tools/harness-eval/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/redhat-community-ai-tools/harness-eval/claude-md"><img src="https://agentmods.dev/badge/instructions/redhat-community-ai-tools/harness-eval/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.01147 | $0.01147 |
| Opus 5.5 | $0.00459 | $0.00459 |
| Sonnet 5.5 | $0.00229 | $0.00229 |
| Haiku 4.5 | $0.00115 | $0.00115 |
Grade A, and why
harness-eval 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 18d ago.
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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
harness-eval
AI agent setup evaluation tool. See README.md for usage, features, and installation.
Development
- Python 3.11+, managed with
uv - Run tests:
uv run pytest - Lint:
uv run ruff check src/ tests/ - Format:
uv run ruff format src/ tests/ - Format check (dry run):
uv run ruff format --check src/ tests/ - Type check:
uv run mypy src/ - See
CONTRIBUTING.mdfor adding rules, plans, and PRs - See
CHANGELOG.mdfor release history
Changelog
Every PR that adds a feature, fixes a bug, or changes behavior must include a CHANGELOG.md entry under [Unreleased]. Add it in the appropriate subsection (Added, Changed, Fixed, Removed). One line per change.
To cut a release: uv run scripts/release.py minor (or patch/major). This moves unreleased entries to a dated version, bumps pyproject.toml, commits, and tags.
Before committing
The CI runs 5 jobs: lint, typecheck, test, security gate, lint gate. All must pass. Run this before committing:
uv run ruff format src/ tests/ && uv run ruff check src/ tests/ && uv run pytest tests/ -q
uv run harness-eval harness-autonomy . # decidable rules only; exit 1 FAIL, 2 REVIEW_REQUIRED
uv run harness-eval harness-security . --fail-on-error # security gate: errors block (warnings are informational)
uv run harness-eval harness-gate . # validated block rules; exits 1 on errors and warnings
uv run harness-eval harness-lint . --all --fail-on-error # every rule: errors block, advisory findings do not
The most common CI failure is forgetting ruff format. The security gate blocks on any security finding (even warnings). harness-autonomy and harness-gate are the integrity gates.
Project structure
src/harness_eval/- main packagecli/- Click CLI package (autonomy.py,gate.py,security.py,lint.py,review.py,doctor.py,baseline.py,rules.py)config/- severity presets (recommended/strict) and the effect-derived rule sets each command runscore/- setup discovery, fingerprinting, component typesdiscoverers/- per-tool discoverer classes (ToolDiscovererABC); add new assistants here
inspection/- static analysis: parsers, lint engine, 92 rules, suppression, auto-fixrules/security/_shared.py- shared scanning logic used by security rules across component typesharness_formats/- per-format mappers (one persource_tool) into the normalized pipeline-harness model consumed byrules/harness/
rubric/- LLM-based issue detection; prompts inrubric/prompts/analysis/- system-level analysis (budget, triggers, dependencies, context utilization)output/- report generation (terminal + JSON)utils/- token counting, TF-IDF similarity, frontmatter parsing, LLM client, path safetydata/- versioned JSON data files (builtins, tautological patterns) for knowledge that decays
skills/- plugin skills with SKILL.md + rubric files + scriptstests/- pytest test suite with fixtures
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.
- 18d ago Changed · +2 lines · +60 tokens per session bae7af2941b2
- 22d ago Changed · +1 tokens per session 9d3ac0b2b64f
- 1mo ago Changed · +1 lines · +36 tokens per session 41398a6b58f8
- 1mo ago First seen · 64 lines · 1,050 tokens per session scan A bb55ff5f92cf
harness-eval CLAUDE.md is an instructions file published in the GitHub repository redhat-community-ai-tools/harness-eval (32 stars, last pushed 17d ago), licensed Apache-2.0. It adds 1,147 tokens to every session, about $0.0046 per session on Opus 5.5. 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 2026-08-30.
Other instructions, from other repositories
data-scientist copilot-instructions.md
Repository instructions for a data-science workflow in GitHub Copilot, covering datasets, statistics, and analysis files.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).