skill-eval

skill-eval is a skill for Claude Code, Codex from joris887/exosuit. It costs 32 tokens per session (1,537 once invoked), scanned A, original, MIT.

A workflow for testing and comparing the effectiveness of coding-agent skills. It can run a skill against a scenario, capture a baseline, inspect metrics, compare versions, and detect regressions.

In plain words
What is it for?
Use it to evaluate a skill with a sample task, compare old and new implementations, record expected results, inspect evaluation criteria, and check for regressions.
Why use it?
It provides a repeatable way to see whether a skill behaves as expected and whether a new version has become worse than an earlier one.

Skill for Claude CodeCodex

Part of the exosuit plugin — 44 skills, 1 command, 9 agents, 10 hooks shipped together

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.

agentmods
npx agentmods add skills/joris887/exosuit/skill-eval
Any agent
npx skills add joris887/exosuit --skill skill-eval
Clone the repo
git clone --depth 1 https://github.com/joris887/exosuit

Made for: Claude Code, Codex.

Or install exosuit, the plugin that ships this one along with the rest of its 44 skills, 1 command, 9 agents, 10 hooks.

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.

agentmods badge for skill-eval

README.md
[![agentmods](https://agentmods.dev/badge/skills/joris887/exosuit/skill-eval.svg)](https://agentmods.dev/skills/joris887/exosuit/skill-eval)
Your own site
<a href="https://agentmods.dev/skills/joris887/exosuit/skill-eval"><img src="https://agentmods.dev/badge/skills/joris887/exosuit/skill-eval.svg" alt="Measured on agentmods" height="20"></a>
Per session 32 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,537 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00032 $0.01537
Opus 5 $0.00016 $0.00768
Sonnet 5 $0.00006 $0.00307
Haiku 4.5 $0.00003 $0.00154

Measured 3d ago against content hash 8c955efe7197, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-eval 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 3d 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/skill-eval/SKILL.md · 202 lines

How it starts

The opening of the file, as written. The whole thing — 202 lines — stays where its author put it; the contents beside it link to each section on GitHub.


skill-eval

Evaluating skill: $ARGUMENTS

Modes

Parse the first argument to determine mode:

Mode Syntax Purpose
eval /skill-eval eval <skill-name> --scenario "<prompt>" Test a skill against a known scenario
compare /skill-eval compare <skill-name> --old <path> --new <path> --scenario "<prompt>" A/B test two versions of a skill
metrics /skill-eval metrics <skill-name> Analyze a skill's evaluation criteria and pressure scenarios
baseline /skill-eval baseline <skill-name> --scenario "<prompt>" Capture baseline output for regression detection
regression /skill-eval regression <skill-name> Compare current skill against saved baseline

Mode: eval

Test a skill against a scenario and grade the output against expectations.

Process

  1. Load the skill — Read .claude/skills/<skill-name>/SKILL.md
  2. Load evaluation criteria — Read the skill's ## Evaluation Criteria section (if present in SKILL.md or SKILL_TEMPLATE.md)
  3. Define the scenario — Use the provided --scenario or prompt the user for one
  4. Execute mentally — Walk through what Claude would do with this skill loaded, given the scenario prompt. Trace the expected decision points, tool calls, and outputs.
  5. Grade against criteria — For each evaluation criterion, determine PASS/FAIL with evidence

Output

## Skill Evaluation: <skill-name>

### Scenario
> <the test scenario>

### Expected Behavior
| # | Criterion | Status | Evidence |
|---|-----------|--------|----------|
| 1 | [criterion text] | PASS/FAIL | [why] |

### Hard Gate Compliance
- [ ] All hard gates would be respected
- [ ] Red flags table would prevent shortcuts (if applicable)

### Context Budget
- Estimated SKILL.md tokens: X
- References loaded: [list]
- Total estimated tokens: X

### Verdict: PASS / FAIL / PARTIAL
[Summary of findings]

Read the full file on GitHub · 202 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

  1. 3d ago First seen · 202 lines · 32 tokens per session scan A 8c955efe7197

Subscribe to this mod's changes

skill-eval is a skill published in the GitHub repository joris887/exosuit (4 stars, last pushed 14d ago), licensed MIT. It adds 32 tokens to every session and 1,537 once invoked, about $0.0002 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.

Related

Other skills, from other repositories

build-teaql-app

Build or change a TeaQL application in Java, Rust, Go, Swift, Python, C#/.NET, or TypeScript, including Kotlin/JVM applications that consume Java-generated libraries. Mandatory order: first draft and save a complete KSML model, then verify the client and evaluate that saved model, repair it through repeated evaluation…

teaql/teaql-agent-kit · 112 tokens

moai-kanban-foreman

One unattended kanban foreman iteration: watch the backlog queue, dispatch the next operator-picked card to an isolated worker, collect completion evidence on read (not on claims), and report. This is the body the project's loop.md driver invokes each iteration of a bare /loop; it can also be invoked directly to test…

modu-ai/moai-adk · 76 tokens

linkedin-post

You take a YouTube video transcript and produce a single LinkedIn post draft.

owainlewis/youtube-tutorials · 0 tokens

orchestrator-lanes

Claude Code dev-orchestrator only. File-based multi-lane PM playbook (score, DAG, run-controller, L0/L1/L2, ship). Use when this session IS that agent, or user says info / справка / lane-stack:orchestrator-lanes info. SKIP: Grok, Codex, Kimi, Qwen, AGY, Cursor writer CLIs and any default coding agent — do not load, do…

VKirill/claude-lane-stack · 104 tokens

source-command-methodology-advisor

Analyzes your codebase and asks 3 targeted questions to recommend the right AI-assisted development methodology stack.

FlorianBruniaux/claude-code-ultimate-guide · 27 tokens

source-command-ccguide-daily

Daily update check — official Anthropic docs diff + guide/CC releases digest.

FlorianBruniaux/claude-code-ultimate-guide · 23 tokens