Stage 2 audit (ARIS §3.1). For each experimental claim, decide supported / partially / invalidated against the logs.
Stage 2 audit (ARIS §3.1). For each experimental claim, decide supported / partially / invalidated against the logs.
npx skills add https://github.com/PAMF2/oniro-colab --skill audit-result-to-claim复制此命令并粘贴到 Claude Code 中以安装该技能
Stage 1 audit (ARIS §3.1). Verify the training run actually converged, seeds were logged, no data leakage.
Stage 3 audit (ARIS §3.1). Fresh zero-context reviewer re-reads wiki narrative; cross-checks against raw results.
Compute novelty score for a candidate descriptor via k-NN distance in archive descriptor space.
On REJECT verdict, write a structured entry to the failure ledger so future executors skip the closed branch.
Read wiki frontier + failure ledger, emit one typed mutation as unified diff. Used by Executor agent.
After a mutation passes (or just falls in an empty novelty cell), update the MAP-Elites archive.
| name | audit-result-to-claim |
| description | Stage 2 audit (ARIS §3.1). For each experimental claim, decide supported / partially / invalidated against the logs. |
| trigger | reviewer-stage-2 |
| allowed_tools | ["read_file","grep","glob"] |
| output | json {verdict, items} |
You are the Reviewer (Stage 2). You hold the run's quantitative claims to the raw logs.
metrics.jsonl / wandb.log /
the wiki variant frontmatter.supported — numbers match within rounding tolerance.partially_supported — claim is true on a subset of the eval set, not all of it.invalidated — no log line supports the claim.supported iff every individual claim is supported.invalidated iff any single claim is invalidated.partially_supported.{
"verdict": "supported" | "partially_supported" | "invalidated",
"items": [
"claim 'ARC-3 dev 0.32' supported by metrics.jsonl line 4812",
"claim 'Gödel gate accept rate 22%' partially_supported: 22% on weeks 1-3, 14% on week 4"
]
}