| name | targeted-revision |
| description | Perform one AutoSciRub targeted revision round using failed executable-rubric criteria and criterion-level verification feedback. Use when Codex needs to update a research artifact by adding missing experiments, comparisons, evidence, analysis, figures, tables, or report text while preserving supported claims and recording a structured revision plan and change summary. |
Targeted Revision
Purpose
Use failed criteria and diagnoses to revise the current artifact once. This implements pi_rev(A_i^{(t)}, Delta_i^{(t)}, E_i | R_i) from AutoSciRub.
This skill performs one bounded revision round. Run verification again only when the user or controller requests it.
All .autoscirub/ paths below refer to the state directory resolved by the controller. When invoked alone, use the user’s override, then AUTOSCIRUB_STATE_DIR, then project config state_dir, then .autoscirub/.
Inputs
Read:
.autoscirub/executable_rubric.json
.autoscirub/verification_report.json or the requested round-specific verification file
- current report and supporting artifacts
- task-visible data/code/tools needed to address gaps
Output
For revision round N, write:
.autoscirub/revisions/round-NNN/revision_plan.json
.autoscirub/revisions/round-NNN/change_summary.json
Use this revision-plan shape:
{
"schema_version": "1.0",
"round": 1,
"targets": [
{
"criterion_id": "C1",
"gap": "...",
"planned_actions": ["..."],
"expected_artifacts": ["..."],
"risk_or_dependency": "..."
}
]
}
Use this change-summary shape:
{
"schema_version": "1.0",
"round": 1,
"changes": [
{
"criterion_id": "C1",
"actions_completed": ["..."],
"artifacts_changed": ["relative/path"],
"remaining_gap": "none or specific limitation"
}
]
}
Procedure
- Select failed criteria from the verification report, prioritizing high-priority evidence gaps.
- Convert each gap into concrete actions: run missing analysis, add comparison, regenerate artifact, strengthen explanation, or remove unsupported claim.
- Update code, results, figures, tables, and report text only as needed to address the failed criteria.
- Preserve existing supported results and claims unless the new evidence contradicts them.
- Record the plan before substantive changes when practical, then record the final change summary.
Rules
- Perform exactly one revision round per invocation.
- Do not broaden the task beyond failed criteria unless a dependency is necessary to satisfy them.
- Do not fabricate results, labels, baselines, or references.
- If a requested fix is infeasible with available data, document the limitation and revise the report to avoid unsupported claims.
- Prefer adding missing evidence over cosmetic report edits.