agent-eval SKILL.md skillA

agent-eval from timurgaleev/vibestack is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add timurgaleev/vibestack/agent-eval@git:20261012.6270fc8

Pins git:20261012.6270fc8 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
timurgaleev/vibestack
Path
skills/agent-eval/SKILL.md
Versions recorded
7
Latest version
git:20261012.6270fc8, committed Oct 12, 2026
First version recorded
Sep 2, 2026
Size
29,230 bytes
Repository stars
7
License
MIT

Build and run an eval harness for an LLM agent or prompt, with scoring and a regression gate.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

## When to invoke

Use when: "evaluate the agent", "build an eval set", "LLM as judge", "prompt regression test", "did the prompt change make it better", "measure prompt quality", "eval harness".

Do not use for web page performance (`/benchmark`) or for comparing several models on a single prompt (`/deliberation:ask-all`, when that plugin is installed). This skill builds a repeatable test suite for one unit of LLM behavior in the user's own project.

## Preamble

```bash
eval "$("${VIBESTACK_HOME:-$HOME/.vibestack}/bin/vibe-slug" 2>/dev/null)" 2>/dev/null || SLUG="unknown"
_LEARN_FILE="${VIBESTACK_HOME:-$HOME/.vibestack}/projects/${SLUG:-unknown}/learnings.jsonl"
if [ -f "$_LEARN_FILE" ]; then
  _LEARN_COUNT=$(wc -l < "$_LEARN_FILE" 2>/dev/null | tr -d ' ')
  echo "LEARNINGS: $_LEARN_COUNT entries loaded"
  if [ "$_LEARN_COUNT" -gt 5 ] 2>/dev/null; then
    "${VIBESTACK_HOME:-$HOME/.vibestack}/bin/vibe-learnings-search" --limit 5 2>/dev/null || true
  fi
else
  echo "LEARNINGS: none yet"
fi
```

{{include lib/snippets/session-host.md}}

{{include lib/snippets/decision-brief.md}}

{{include lib/snippets/working-protocols.md}}

{{include lib/snippets/state-protocols.md}}
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add timurgaleev/vibestack/agent-eval@sha256:0a0795bda26da8d5

The line at the top of this page pins the label git:20261012.6270fc8, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_lwgukgdqj6slp2vu.svg)](https://markdownregistry.com/a/art_lwgukgdqj6slp2vu)

3 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 7 versions of this file, between Sep 2, 2026 and Oct 12, 2026, about one every 7 days.

versioncommittedcommitsizeaudit
git:20261012.6270fc8 latest2026-10-12 6270fc8 29,230 BA view · diff
git:20261012.0f258f32026-10-12 0f258f3 29,086 BA view · diff
git:20261011.be543ef2026-10-11 be543ef 28,319 BA view · diff
git:20261009.553d4132026-10-09 553d413 27,435 BA view · diff
git:20261009.5df8a1a2026-10-09 5df8a1a 27,431 BA view · diff
git:20260902.2f747e92026-09-02 2f747e9 27,595 BA view · diff
git:20260902.281273b2026-09-02 281273b 21,077 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (29230 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

timurgaleev/vibestack · 7 stars · license MIT · pushed 2026-10-15 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_lwgukgdqj6slp2vu
GET https://markdownregistry.com/api/v1/resolve?ref=timurgaleev/vibestack/agent-eval
GET https://markdownregistry.com/api/v1/blob/0a0795bda26da8d512bb236170cbbd33b6a35d66759ec3084bbc4ce3c7a2d92c

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from timurgaleev/vibestack

CLAUDE.md CLAUDE.md file
timurgaleev/vibestack · CLAUDE.md
git:20261013.0979e18 · audit A · 7 stars
CLAUDE.md@config/claude CLAUDE.md file
timurgaleev/vibestack · config/claude/CLAUDE.md
git:20261009.16b4f3c · audit A · 7 stars
AGENTS.md@config/codex AGENTS.md file
timurgaleev/vibestack · config/codex/AGENTS.md
git:20261013.e98bb89 · audit B · 7 stars
address-pr-review SKILL.md skill
timurgaleev/vibestack · skills/address-pr-review/SKILL.md · Work a PR's open review threads and failing CI checks to a close: fix or explain, push, reply, resolve.
git:20261011.7ad9aa7 · audit A · 7 stars
ai-cost-guard SKILL.md skill
timurgaleev/vibestack · skills/ai-cost-guard/SKILL.md · Find code paths that can run up an unbounded LLM or inference bill and require a dollar cap before shipping.
git:20261011.be543ef · audit A · 7 stars
autoplan SKILL.md skill
timurgaleev/vibestack · skills/autoplan/SKILL.md · Run the CEO, design, eng and DX plan reviews in one pass and surface only the taste decisions for approval.
git:20261009.e3b2bd3 · audit A · 7 stars
aws-cost SKILL.md skill
timurgaleev/vibestack · skills/aws-cost/SKILL.md · Read-only AWS cost review: where last month's money went, what changed, and the top actions to cut the bill.
git:20261009.553d413 · audit A · 7 stars
bedrock-guardrails SKILL.md skill
timurgaleev/vibestack · skills/bedrock-guardrails/SKILL.md · Read-only audit of Amazon Bedrock guardrails, IAM, logging and data residency before shipping an LLM feature.
git:20261009.553d413 · audit A · 7 stars
benchmark SKILL.md skill
timurgaleev/vibestack · skills/benchmark/SKILL.md · Detect page performance regressions: baseline load times, Core Web Vitals and bundle sizes, compare per PR.
git:20261009.553d413 · audit A · 7 stars
browse SKILL.md skill
timurgaleev/vibestack · skills/browse/SKILL.md · Fast headless browser for QA and dogfooding: navigate, click, fill forms, screenshot and assert page state.
git:20261009.553d413 · audit F · 7 stars
canary SKILL.md skill
timurgaleev/vibestack · skills/canary/SKILL.md · Post-deploy canary monitoring: watch the live app for console errors, slowdowns and page failures.
git:20261012.eeb894c · audit B · 7 stars
careful SKILL.md skill
timurgaleev/vibestack · skills/careful/SKILL.md · Safety guardrails that warn before destructive commands like rm -rf, DROP TABLE, force-push or reset --hard.
git:20261015.7f2ac1c · audit A · 7 stars

Every file in timurgaleev/vibestack

Other files named agent-eval

agent-eval SKILL.md skill
affaan-m/ecc · docs/ja-JP/skills/agent-eval/SKILL.md · カスタムタスクでコーディングエージェント(Claude Code、Aider、Codex など)をヘッドツーヘッドで比較し、合格率、コスト、時間、一貫性のメトリクスを測定します
git:20260517.ec9ace9 · audit A · 284,510 stars
agent-eval SKILL.md skill
ericrisco/rsc-harness · skills/agent-eval/SKILL.md · Use when measuring whether an LLM or agent system actually got better and gating merges on it: golden sets, fixing an…
git:20260729.3ecb845 · audit A · 236 stars
agent-eval SKILL.md skill
wankimmy/bossku-superpowers · skills/bosskuai-agent-benchmark-comparison/references/upstream/SKILL.md · Head-to-head comparison of coding agents (Claude Code, Aider, Codex, etc.) on custom tasks with pass rate, cost, time…
git:20261012.feb2361 · audit A · 22 stars
agent-eval SKILL.md skill
ethanyoq/skill-hub · archives/user-installed-skills/agent-eval/61ca5bbbd5c736b7/agent-eval/SKILL.md · Benchmark CodeGraph retrieval quality on a real codebase by comparing agent behavior with vs without CodeGraph. Use…
git:20260912.84c6c93 · audit A · 11 stars

Browse by kind, by grade A, or by owner.