eval-agents

eval-agents is a skill for Claude Code, Codex from me2resh/apexyard. It costs 44 tokens per session (6,772 once invoked), scanned A, original, MIT.

An evaluation tool for measuring how well code-review agents find known defects in previously reviewed pull requests.

In plain words
What is it for?
Use it to test agents such as Rex, Hakim, or Tariq on a labeled pull-request collection and measure their review accuracy against a pass/fail threshold.
Why use it?
It replaces subjective ratings of review wording with comparisons against a fixed set of confirmed defects. It reports missed defects, false positives, and how often approvals were correct.

Skill for Claude CodeCodex

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/me2resh/apexyard/eval-agents
Any agent
npx skills add me2resh/apexyard --skill eval-agents
Clone the repo
git clone --depth 1 https://github.com/me2resh/apexyard

Made for: Claude Code, Codex.

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 eval-agents

README.md
[![agentmods](https://agentmods.dev/badge/skills/me2resh/apexyard/eval-agents.svg)](https://agentmods.dev/skills/me2resh/apexyard/eval-agents)
Your own site
<a href="https://agentmods.dev/skills/me2resh/apexyard/eval-agents"><img src="https://agentmods.dev/badge/skills/me2resh/apexyard/eval-agents.svg" alt="Measured on agentmods" height="20"></a>
Per session 44 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,772 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 $0.00044 $0.06772
Opus 5 $0.00022 $0.03386
Sonnet 5 $0.00009 $0.01354
Haiku 4.5 $0.00004 $0.00677

Measured 4d ago against content hash 77ab35775b39, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eval-agents 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 4d ago.

The scan reads SKILL.md. This mod also ships 4 executable files (lib/snapshot-diff.sh, lib/validate-corpus.sh, tests/smoke.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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/eval-agents/SKILL.md · 394 lines

How it starts

The opening of the file, as written. The whole thing — 394 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/eval-agents — Review-Agent Eval Harness

Scores one of the framework's review agents (Rex / Hakim / Tariq) against a labeled corpus of real, already-reviewed PRs with frozen ground-truth defect sets. Reports catch-rate, false-positive-rate, and approve-precision (the headline metric) against a configurable pass/fail threshold.

Why this isn't an LLM-judge rating review prose

Spike #825 tested that shape first — score a review agent's text output against a 0-4 rubric, calibrated against recorded verdicts — and found it at chance on the one question that matters: was an approval justified? A fluent, verification-heavy wrong approval scored as well as or better than genuinely correct approvals (docs/spike-825/findings.md). This skill instead:

  1. Never asks a judge to rate review text. Ground truth is a frozen set of real defects, established once, offline, by a human, from actual re-review disagreements and confirmed fixes — never re-derived at run time, never established by the agent being measured.
  2. Runs the agent-under-test fresh against the corpus entry's diff, and mechanically/semantically compares its findings to the frozen defect set: caught / missed / false-alarm.
  3. Reports approve-precision as the headline metric — the rate at which the agent's approvals were actually justified. A missed BLOCKING/HIGH defect is an automatic WARN regardless of the aggregate score.

Full rationale: AgDR-0089. Corpus format: docs/eval-agents/SCHEMA.md.

Usage

/eval-agents rex                                            # run against the seeded starter corpus
/eval-agents hakim --corpus docs/eval-agents/corpus/hakim.json
/eval-agents tariq --corpus my-corpus.json                  # tariq has no starter corpus — required
/eval-agents rex --check-only                                # validate corpus schema only, no spawns

<agent> is one of rex, hakim, tariq — the three review agents that produce a verdict over a diff and share the APPROVED / CHANGES REQUESTED / COMMENT vocabulary. Naqid is out of scope for v1 (it challenges premises, not diffs — no defect-set structure to score against; see AgDR-0089 § Decision point 7).

Read the full file on GitHub · 394 lines

Files

What ships with it

4 files 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. 4d ago First seen · 394 lines · 44 tokens per session scan A 77ab35775b39

Subscribe to this mod's changes

eval-agents is a skill published in the GitHub repository me2resh/apexyard (498 stars, last pushed today), licensed MIT. It adds 44 tokens to every session and 6,772 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 2026-08-30.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens

chat-perf

Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

microsoft/vscode · 51 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens