AGENT SKILL

agent-eval

Head-to-head comparison of coding agents (Claude Code, Aider, Codex, etc.) on custom tasks with pass rate, cost, time, and consistency metrics

License 许可
MIT
最近更新
2026-09-24
Stars
★ 0
Forks
1

01 前言 Front Matter

name
agent-eval
tools
Read, Write, Edit, Bash, Grep, Glob
license
MIT
metadata
{ "origin": "ECC" }
description
Head-to-head comparison of coding agents (Claude Code, Aider, Codex, etc.) on custom tasks with pass rate, cost, time, and consistency metrics

03 SKILL.md 正文

Agent Eval Skill

A lightweight CLI tool for comparing coding agents head-to-head on reproducible tasks. Every "which coding agent is best?" comparison runs on vibes — this tool systematizes it.

When to Activate

  • Comparing coding agents (Claude Code, Aider, Codex, etc.) on your own codebase
  • Measuring agent performance before adopting a new tool or model
  • Running regression checks when an agent updates its model or tooling
  • Producing data-backed agent selection decisions for a team

Installation

Note: Install agent-eval from its repository after reviewing the source.

Core Concepts

YAML Task Definitions

Define tasks declaratively. Each task specifies what to do, which files to touch, and how to judge success:

name: add-retry-logic
description: Add exponential backoff retry to the HTTP client
repo: ./my-project
files:
  - src/http_client.py
prompt: |
  Add retry logic with exponential backoff to all HTTP requests.
  Max 3 retries. Initial delay 1s, max delay 30s.
judge:
  - type: pytest
    command: pytest tests/test_http_client.py -v
  - type: grep
    pattern: "exponential_backoff|retry"
    files: src/http_client.py
commit: "abc1234"  # pin to specific commit for reproducibility

Git Worktree Isolation

Each agent run gets its own git worktree — no Docker required. This provides reproducibility isolation so agents cannot interfere with each other or corrupt the base repo.

Metrics Collected

Metric What It Measures
Pass rate Did the agent produce code that passes the judge?
Cost API spend per task (when available)
Time Wall-clock seconds to completion
Consistency Pass rate across repeated runs (e.g., 3/3 = 100%)

Workflow

1. Define Tasks

Create a tasks/ directory with YAML files, one per task:

mkdir tasks
# Write task definitions (see template above)

2. Run Agents

Execute agents against your tasks:

agent-eval run --task tasks/add-retry-logic.yaml --agent claude-code --agent aider --runs 3

Each run:

  1. Creates a fresh git worktree from the specified commit
  2. Hands the prompt to the agent
  3. Runs the judge criteria
  4. Records pass/fail, cost, and time

3. Compare Results

Generate a comparison report:

agent-eval report --format table
Task: add-retry-logic (3 runs each)
┌──────────────┬───────────┬────────┬────────┬─────────────┐
│ Agent        │ Pass Rate │ Cost   │ Time   │ Consistency │
├──────────────┼───────────┼────────┼────────┼─────────────┤
│ claude-code  │ 3/3       │ $0.12  │ 45s    │ 100%        │
│ aider        │ 2/3       │ $0.08  │ 38s    │  67%        │
└──────────────┴───────────┴────────┴────────┴─────────────┘

Judge Types

Code-Based (deterministic)

judge:
  - type: pytest
    command: pytest tests/ -v
  - type: command
    command: npm run build

Pattern-Based

judge:
  - type: grep
    pattern: "class.*Retry"
    files: src/**/*.py

Model-Based (LLM-as-judge)

judge:
  - type: llm
    prompt: |
      Does this implementation correctly handle exponential backoff?
      Check for: max retries, increasing delays, jitter.

Best Practices

  • Start with 3-5 tasks that represent your real workload, not toy examples
  • Run at least 3 trials per agent to capture variance — agents are non-deterministic
  • Pin the commit in your task YAML so results are reproducible across days/weeks
  • Include at least one deterministic judge (tests, build) per task — LLM judges add noise
  • Track cost alongside pass rate — a 95% agent at 10x the cost may not be the right choice
  • Version your task definitions — they are test fixtures, treat them as code

Links

04 同仓库更多 Skills

NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

ml-best-practices

CRITICAL RULE: You MUST use this skill whenever the task involves any machine learning tasks or data analysis. Use this skill if the user's prompt or requirements mention any of the following: * Clustering * Classification * Regression * Time series forecasting * Statistical testing * Model comparison * ML * Data analysis SQL/BigQuery ML HANDOFF: If the user requires a SQL solution, use this skill to dictate the ANALYSIS STEPS (e.g., markdown analysis cells, visualization logic), but defer to `bigquery` for all SQL syntax.

★ 0
NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

verification-loop

A comprehensive verification system for Claude Code sessions.

★ 0
NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

returns-reverse-logistics

Codified expertise for returns authorization, receipt and inspection, disposition decisions, refund processing, fraud detection, and warranty claims management. Informed by returns operations managers with 15+ years experience. Includes grading frameworks, disposition economics, fraud pattern recognition, and vendor recovery processes. Use when handling product returns, reverse logistics, refund decisions, return fraud detection, or warranty claims.

★ 0
NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

recursive-decision-ledger

Use when the user asks for repeated rollouts, marked decision processes, high-dimensional search, stochastic optimization, local-optima exploration, ensemble comparison, or recursive reasoning with a visible evidence trail.

★ 0
NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

quality-nonconformance

Codified expertise for quality control, non-conformance investigation, root cause analysis, corrective action, and supplier quality management in regulated manufacturing. Informed by quality engineers with 15+ years experience across FDA, IATF 16949, and AS9100 environments. Includes NCR lifecycle management, CAPA systems, SPC interpretation, and audit methodology. Use when investigating non-conformances, performing root cause analysis, managing CAPAs, interpreting SPC data, or handling supplier quality issues.

★ 0
NguyenPhuocAnhDung
NguyenPhuocAnhDung/oil_forecast_tail_risk

production-scheduling

Codified expertise for production scheduling, job sequencing, line balancing, changeover optimization, and bottleneck resolution in discrete and batch manufacturing. Informed by production schedulers with 15+ years experience. Includes TOC/drum-buffer-rope, SMED, OEE analysis, disruption response frameworks, and ERP/MES interaction patterns. Use when scheduling production, resolving bottlenecks, optimizing changeovers, responding to disruptions, or balancing manufacturing lines.

★ 0