Skip to content

openJiuwen-ai/sciencediscovery / skills/computation-reviewer

computation-reviewer

Verified

Check whether numeric claims marked with Evidence aliases in an Artifact agree with the linked Evidence content and context.

FileKindSize
SKILL.mdskill md1.6 KB
---
name: computation-reviewer
description: Check whether numeric claims marked with Evidence aliases in an Artifact agree with the linked Evidence content and context.
---

# Computation Reviewer

Run on the same locked Artifact with its supplied Evidence Bundle. In the bounded Deep
pipeline, this local Evidence comparison runs before Citation source verification.

## Required checks

1. Review exactly one supplied numeric Claim and its matching Evidence Bundle entry per run. The
   Claim must contain a number and an `[evN]` alias in the same sentence or Markdown table cell.
2. Compare value, unit, range, sample size, population, subgroup, endpoint, and statistical
   context with the linked Evidence.
3. Distinguish exact support, value or unit contradiction, scope mismatch, interpretation
   overreach, and insufficient evidence.
4. Return `INCONCLUSIVE` when the Evidence excerpt does not contain enough context; do not report
   insufficient content as a value mismatch.

## Output

Return only the required structured result. Use a concrete `COMPUTATION_*` finding code such as
`COMPUTATION_EVIDENCE_VALUE_MISMATCH`, `COMPUTATION_EVIDENCE_SCOPE_MISMATCH`,
`COMPUTATION_EVIDENCE_INTERPRETATION_OVERREACH`, or `COMPUTATION_EVIDENCE_INSUFFICIENT`.
Preserve the supplied Evidence alias and provide one concise reason; do not invent a separate
correction field that the review protocol cannot persist.

## Boundaries

- Do not recompute results, execute code, modify Artifacts, or create another agent.
- Do not treat topical similarity as numeric support.
- Preserve Evidence aliases and provide a concise reason for every mismatch.
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 322kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 322kUpdated Claude CodeCodexVerifiedInstall