agent-eval

A framework for testing and comparing AI assistants. An evaluation is a structured test of how well an assistant handles example tasks, using measures such as correctness or consistency.

In plain words
What is it for?
Use it to run single- or multi-turn evaluations, compare assistant versions, analyze failures, publish benchmark results, and optimize prompts.
Why use it?
It replaces informal testing with repeatable benchmark runs, metric-based grading, comparisons, and failure analysis.

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/googlecloudplatform/professional-services/agent-eval
Any agent
npx skills add GoogleCloudPlatform/professional-services --skill agent-eval
Clone the repo
git clone --depth 1 https://github.com/GoogleCloudPlatform/professional-services

Made for: Claude Code, Codex.

Per session 136 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,505 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 $0.00136 $0.01505
Opus 5 $0.00068 $0.00753
Sonnet 5 $0.00027 $0.00301
Haiku 4.5 $0.00014 $0.00151

Measured 2d ago against content hash f7c90a219c3e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-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 2d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/parse_eval_summary.py), 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.

tools/agent-eval/skills/agent-eval/SKILL.md · 109 lines

How it starts

The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Evaluation & Continuous Optimization Framework

This skill defines the end-to-end evaluation, benchmarking, automated optimization, and publication workflow using the agent-eval CLI Pipeline aligned with google/agents-cli and the ADK Quality Flywheel (adk.dev/optimize).


1. Reference Architecture & Deep Guides

Reference Guide Contents
references/dataset_schema.md Canonical dataset.jsonl schema (single-turn, multi-turn, multi-agent topologies).
references/metrics_guide.md Declarative eval_config.yaml specification across 6 metric kinds.
references/gepa_optimization.md Automated genetic prompt evolution via GEPARootAgentPromptOptimizer.
references/failure_triage.md 2-Tier loss clustering and Context Engineering remediation strategies.

2. The Standard agent-eval run Benchmark Command

Always execute benchmark sweeps using the standardized --feature, --tag, and --publish taxonomy against an active API server endpoint (--base-url):

export AGENT_EVAL_NO_PAUSES=1
export GOOGLE_GENAI_USE_VERTEXAI=1
export GOOGLE_CLOUD_PROJECT=<PROJECT_ID>

agent-eval run \
  --agent-dir app \
  --base-url http://localhost:8080 \
  --feature "<feature_or_branch_name>" \
  --tag "<short_iteration_tag>" \
  --description "<one-line summary of changes tested>" \
  --sim-parallelism 6 \
  --publish \
  --compare-to "<baseline_run_id_or_path>"

Key CLI Flags & Defaults

  • --feature: Git feature branch or capability under test (defaults to active git branch).
  • --tag: Concise iteration slug (e.g. direct-bypass-v1, calibrated-prompt-v2).
  • --description: Human-readable context saved into eval_summary.json and rendered in the dashboard.
  • --sim-parallelism 6: Runs 6 scenarios in parallel, cutting multi-turn sweeps down to ~2.5 minutes.
  • --publish: Automatically syncs the entire output run directory to Google Cloud Storage (gs://<PROJECT_ID>-eval-artifacts/runs/<run_id>/).
  • --compare-to: Resolves a baseline run (locally or directly from GCS) and generates delta percentage scorecards.

Read the full file on GitHub · 109 lines

Files

What ships with it

5 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. 2d ago First seen · 109 lines · 136 tokens per session scan A f7c90a219c3e

Subscribe to this mod's changes

agent-eval is a skill published in the GitHub repository GoogleCloudPlatform/professional-services (3,065 stars, last pushed 11d ago), licensed Apache-2.0. It adds 136 tokens to every session and 1,505 once invoked, about $0.0007 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

review-prs

Review a GitHub pull request in the googleapis/mcp-toolbox repo against the team's reviewer checklist: PR title/description conventions, linked issue, logic errors and unhandled edge cases, breaking changes, test coverage, docs updates, security (input handling), and new dependencies. Use whenever a maintainer asks…

googleapis/mcp-toolbox · 162 tokens

fix-failing-tests

Diagnose a failing test in the googleapis/mcp-toolbox repo and land a fix by reasoning from the actual error: read the failure, reproduce it, shrink it until the cause is forced into the open, then fix the cause. Use this whenever a test or CI job is red, a build breaks after a change, many packages fail at once, or a…

googleapis/mcp-toolbox · 87 tokens

stale-sweep

Sweep the googleapis/mcp-toolbox repo for issues and PRs with no real activity in N days (default 60), sort each by whose silence it is (the author's, ours, or nobody's), and draft the nudge or close comment. Use whenever a maintainer asks for a stale sweep, backlog cleanup, or an SLO check, e.g. "stale sweep", "find…

googleapis/mcp-toolbox · 159 tokens

triage-issues

Triage GitHub issues in the googleapis/mcp-toolbox repo: propose the correct labels (type / priority / product / status), check for duplicates, verify a bug has enough info to act on, and draft a triage comment. Use whenever a maintainer asks you to triage, label, categorize, prioritize, or "look at" an issue (or a…

googleapis/mcp-toolbox · 164 tokens

data-parity

Validate that two tables or query results are identical — or diagnose exactly how they differ. Discover schema, identify keys, profile cheaply, then diff. Use for migration validation, ETL regression, and query refactor verification.

AltimateAI/altimate-code · 48 tokens

dbt-develop

REQUIRED before writing or modifying ANY dbt model. Invoke this skill FIRST whenever a task says "create", "build", "add", "modify", "update", "fix", or "refactor" a dbt model, staging file, mart, incremental, or snapshot. Skipping this skill is the leading cause of silent-correctness bugs — models that compile and…

AltimateAI/altimate-code · 294 tokens