A test-running guide for the cabineteer project’s evaluation suite. The suite checks tool handlers directly, without using the network transport normally used to connect to an MCP server.

Running the full eval suite, selecting particular scenarios or categories, listing available tests, viewing verbose results, and producing JSON output for scripts or CI.

It provides a quick way to verify changes and diagnose failures, with filters for tags, difficulty, and scenario names. The documented baseline is 305 scenarios and 1,139 assertions, all passing.

run-evals — Skill for Claude Code by chemrich

Install

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.

Any agent

npx skills add chemrich/cabineteer --skill run-evals

Clone the repo

git clone --depth 1 https://github.com/chemrich/cabineteer

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/chemrich/cabineteer/run-evals/github.svg)](https://agentmods.dev/skills/chemrich/cabineteer/run-evals)

Your own site

<a href="https://agentmods.dev/skills/chemrich/cabineteer/run-evals"><img src="https://agentmods.dev/badge/skills/chemrich/cabineteer/run-evals/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/chemrich/cabineteer/run-evals"><img src="https://agentmods.dev/badge/skills/chemrich/cabineteer/run-evals.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 79 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 666 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00079	$0.00666
Opus 5	$0.00039	$0.00333
Sonnet 5	$0.00016	$0.00133
Haiku 4.5	$0.00008	$0.00067

Measured 10d ago against content hash d2a29ac1fd65, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

run-evals 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 10d 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.

.claude/skills/run-evals/SKILL.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Running the eval suite

The evals are the fast, transport-free integration check for this repo. They import server handler functions directly via TOOL_DISPATCH, so the whole suite runs in about one second. Run them after any non-trivial change — CLAUDE.md treats this as required.

Commands

uv run python -m evals                 # full suite
uv run python -m evals --tag kitchen   # one tag
uv run python -m evals --tag drawer --tag door   # OR across tags (repeatable)
uv run python -m evals --difficulty basic        # basic | standard | advanced
uv run python -m evals --name overflow_drawer_stack   # one scenario (repeatable)
uv run python -m evals --json          # machine-readable (CI/scripting)
uv run python -m evals --verbose       # show passing assertions too
uv run python -m evals --list          # print the catalogue without running

--tag is validated against the auto-derived ALL_TAGS; a typo exits with code 2, and a filter combination that matches zero scenarios exits 1 (both were vacuous-green holes before PR #21 — do not "fix" them back).

Baseline

Green is 305 scenarios / 1139 assertions / 100%. Exit code 0 on all-pass, 1 on any failure. If your change moves the assertion count, that's expected only when you added/removed assertions — otherwise investigate.

Reading a failure

Each failing scenario prints [FAIL] <name> (passed/total) followed by the failing assertion's path and operator. The path is a dot-walk into the tool's JSON result (e.g. summary.errors, exterior.width_mm). To reproduce a single failure in isolation:

uv run python -m evals --name <scenario_name> --verbose

To inspect the actual tool output the assertion walked, drive the handler directly (see the drive-mcp-handlers skill) with the same args from evals/scenarios.py.

Where things live

Harness + assertion operators: evals/harness.py (evaluate_assertion, run_all, TOOL_DISPATCH).
Scenarios (declarative data): evals/scenarios.py.
CLI/flags: evals/__main__.py.

Read the full file on GitHub · 48 lines

Changes

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.

10d ago First seen · 48 lines · 79 tokens per session scan A d2a29ac1fd65

Subscribe to this mod's changes

run-evals is a skill published in the GitHub repository chemrich/cabineteer (3 stars, last pushed 4d ago), licensed Apache-2.0. It adds 79 tokens to every session and 666 once invoked, about $0.0004 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.

Testing

research-engineer

An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.

davila7/claude-code-templates · 43 tokens

tika-eval-compare

Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".

apache/tika · 50 tokens

neuron-evaluation-engineer

Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…

neuron-core/neuron-ai · 77 tokens

run-evals