Skip to content

feat(self-review): add reviewer-team consistency category K to 10-category screen#32

Merged
Yoojin-nam merged 1 commit into
mainfrom
feat/tier1-PR7-self-review-reviewer-team-consistency
May 23, 2026
Merged

feat(self-review): add reviewer-team consistency category K to 10-category screen#32
Yoojin-nam merged 1 commit into
mainfrom
feat/tier1-PR7-self-review-reviewer-team-consistency

Conversation

@Yoojin-nam

Copy link
Copy Markdown
Contributor

PR T1-7: self-review reviewer-team consistency check

Title: feat(self-review): add reviewer-team consistency category K to 10-category screen

Summary

Adds scripts/check_reviewer_team_consistency.py and a new self-review
category K. Reviewer-team consistency (SR/MA-only, fabrication-grade).

Detects manuscripts that simultaneously claim dual independent reviewers
(Methods + PROSPERO) and confess to single-reviewer execution
(Discussion §Limitations). Either claim alone is fine; the conjunction
is read by reviewers as fabrication-grade and causes rejection.

Motivation

Cross-project precedent (anonymized): a reporting-quality SR-of-AI-tools
manuscript had:

  • Methods: "Two reviewers independently screened titles and abstracts ..."
  • PROSPERO record: "Two independent reviewers will perform full-text
    screening and data extraction."
  • Discussion §Limitations: "Single primary reviewer; a 20% sample by an
    additional reviewer is deferred to before submission."

The conjunction admits in Limitations what Methods denies in narrative
form. Fabrication-grade red flag.

Changes

  • skills/self-review/scripts/check_reviewer_team_consistency.py
    section-aware grep for DUAL patterns (Methods + PROSPERO) vs SINGLE
    patterns (Discussion + Limitations).
  • skills/self-review/SKILL.md:
    • new category K with critical-tier check entry
    • command snippet showing how to invoke the script
    • resolution guidance (honest Methods rewrite vs Limitations rewrite)
    • K added to Research-Type Adaptation table (Full for Meta-Analysis,
      N/A elsewhere)
  • skills/self-review/tests/test_reviewer_team_consistency.sh — 4 cases
    (clean / Methods+limits / PROSPERO+limits / single-only).

Test plan

  • bash skills/self-review/tests/test_reviewer_team_consistency.sh
    4/4 PASS
  • bash scripts/validate_skills.sh
  • Manual integration in a self-review walkthrough

Related

  • PR T1-2 (/sync-submission Phase 6 scope drift) — complementary
    detection family
  • PRISMA-DTA reporting checks (existing /check-reporting integration)

🤖 Generated with Claude Code

Adds scripts/check_reviewer_team_consistency.py and a new category K
("Reviewer-team consistency", SR/MA-only, fabrication-grade) to
self-review.

The check detects the conjunction of:
- DUAL claim (Methods or PROSPERO: "two reviewers independently",
  "dual extractors", etc.), AND
- SINGLE confession (Discussion/Limitations: "single primary reviewer",
  "20% sample", "deferred to before submission", "due to resource
  constraints").

Either side alone is fine. Both together is read as fabrication-grade
and ends the round. Provides resolution guidance (honest Methods
rewrite OR Limitations rewrite).

Synthetic fixture tests cover clean / Methods+limits / PROSPERO+limits
/ single-only. All 4 pass.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
@Yoojin-nam Yoojin-nam merged commit 3a0f42f into main May 23, 2026
1 check passed
@Yoojin-nam Yoojin-nam deleted the feat/tier1-PR7-self-review-reviewer-team-consistency branch May 23, 2026 06:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant