Saved reviews
1,665
614 unique PRs
Overview
Review activity and agent performance at a glance.
Accepted findings
2,437
Bot verified; not human confirmed
Addressed after review
179
30 partially addressed
Request records this month
19,832
Ledger entries, including retries
Agent leaderboard
Ranked by verifier-accepted findings across saved reviews. Unique means one contributing agent; shared findings count for each contributor. Different agents may review different PRs.
| Agent / model | Accepted | Unique | Shared | Completion |
|---|---|---|---|---|
| adversarial_claudeclaude-opus-5-5 | 258 | 208 | 96.0% | |
| designgpt-6-luna | 137 | 192 | 98.5% | |
| adversarialgpt-6-astra | 56 | 147 | 100.0% | |
| adversarialgpt-6.1-sol | 73 | 111 | 97.7% | |
| public_contractgpt-6.1-sol | 9 | 124 | 99.2% | |
| dead_codegpt-6-luna | 26 | 35 | 98.7% | |
| concurrencygpt-6-luna | 7 | 47 | 97.6% | |
| testsgpt-6-luna | 6 | 44 | 98.3% | |
| adversarial_geminigoogle/gemini-4-argon | 16 | 31 | 94.3% | |
| independentgpt-6-luna | 2 | 36 | 97.6% | |
| scriptgpt-6.1-sol | 5 | 24 | 96.8% | |
| stategpt-6-luna | 3 | 25 | 98.3% | |
| buildgpt-6-luna | 4 | 19 | 94.7% | |
| challengergpt-6.1-sol | 11 | 2 | 97.7% | |
| verifiergpt-6-luna | 11 | 0 | 93.5% | |
| adversarial_deepseekdeepseek/deepseek-v4-pro-0813 | 0 | 0 | 38.3% | |
| routergpt-6-luna | 0 | 0 | 99.6% | |
| public_contractgpt-6.1-terra | 0 | 0 | 100.0% | |
| major_verifiergpt-6.1-sol | 0 | 0 | 94.1% | |
| collatorgpt-6-luna | 0 | 0 | 98.0% | |
| feedbackgpt-6-luna | 0 | 0 | 87.3% | |
| write_verifiergpt-6.1-sol | 0 | 0 | 52.2% | |
| routergpt-6.1-sol | 0 | 0 | 100.0% | |
| verifierclaude-opus-5-5 | 0 | 0 | 93.0% |
Adversarial model comparison
Observed reviews with completed Codex and Claude (historically Gemini) adversarial stages only; this is not causal recall.
Completed reviews: 361
| Model group | Sole accepted | Shared accepted |
|---|---|---|
| adversarial | 119 | 250 |
| adversarial_claude | 258 | 208 |
| adversarial_gemini | 14 | 30 |