trapstreet.run
H4 for AI workflows · community evaluation harness
200
Tasks
hand-curated, frozen, replayable
2,000
Runs
10 tools × 200 tasks
42
Fabrications caught
trap probes + LLM judge
>Quick start
one-liner · install the trapstreet-eval skill
# 30 seconds, no API key, runs against your current Claude session
$
curl -fsSL https://trapstreet.run/install.sh | bash# then in any Claude Code session:
›
/trapstreet-evalTrack · Resume Tailoring
The leaderboard.
| # | Tool | Tier | Score | Fabrications | $/task | Latency | Pricing |
|---|---|---|---|---|---|---|---|
| 01 | open-resume-tailor indie / open source | GOLD | 84.2 | 0 | $0.008 | 1.4s | Free |
| 02 | Claude (direct prompt) Anthropic | GOLD | 79.1 | 1 | $0.015 | 2.1s | API |
| 03 | Rezi Rezi.ai | GOLD | 71.3 | 6 | $0.042 | 3.8s | $29/mo |
| 04 | Huntr Huntr.co | GOLD | 68.4 | 4 | $0.038 | 2.9s | $39/mo |
| 05 | Kickresume Kickresume | GOLD | 64.0 | 3 | $0.031 | 3.2s | $19/mo |
| 06 | WeekendHack v3.2 indie | BRONZE | 64.0 | 0 | $0.004 | 0.9s | Free |
| 07 | Teal AI Teal HQ | SILVER | 62.8 | 2 | $0.022 | 2.4s | $9/mo |
| 08 | Gemini 2.0 (direct) Google | GOLD | 60.2 | 5 | $0.011 | 1.8s | API |
| 09 | GPT-4o (direct) OpenAI | GOLD | 58.9 | 7 | $0.018 | 2.5s | API |
| 10 | 简历优化助手 Pro (redacted) | GOLD | 41.2 | 14 | $0.091 | 5.2s | ¥199/mo |
Featured trap probe · T-0047 · Quanta Robotics · Trap Street Wall
Find the fakes. We plant verifiable truths inside real tasks. Tools that fabricate trip the trap and land on the public Wall.Read the manifesto