Review stable chapter or manuscript drafts for paragraph flow, transitions, argument continuity, and repeated rhetorical functions before editing. Use after the author-approved spine is stable; route structural drift back to thesis-control.
Scanned 9/3/2026
Install to Claude Code
npx -y skills add yha9806/academic-writing-toolkit --skill logic-review --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Logic Review?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/yha9806-logic-review-academic-writing-toolkit)More formats (shields.io, HTML) on the badges page.
---
name: logic-review
description: Review stable chapter or manuscript drafts for paragraph flow, transitions, argument continuity, and repeated rhetorical functions before editing. Use after the author-approved spine is stable; route structural drift back to thesis-control.
allowed-tools: Read, Glob, Bash, Edit
---
# /logic-review — Paragraph Logic Review
## Purpose
Find paragraphs that may need a topic sentence, a clearer transition, merging,
re-ordering, or removal of duplicated argumentative work. The deterministic
script flags mechanical candidates; the agent separately reviews rhetorical
functions without redefining the manuscript spine.
## Trigger Words
This skill activates on: `logic review`, `flow check`, `paragraph logic`, `/logic-review`.
## Workflow
1. Resolve the bundled helper at `scripts/audit-logic.py` relative to this `SKILL.md`, then run it from the project root:
`python3 {skill_dir}/scripts/audit-logic.py --base-dir . --json`
2. Confirm that the research object, core question, intended use, primary
experiment, headline claim, and evidence boundary are stable under the
active `/thesis-control` contract. If they are not, stop the language pass
and route the issue to drift control or revision escalation.
3. Read the flagged paragraphs and nearby context.
4. Label each relevant sentence or paragraph by its main function: `problem`,
`gap`, `method`, `evidence`, `interpretation`, `boundary`, or `transition`.
5. Flag duplicated functions and monotonous sequences, especially repeated
`evidence -> disclaimer` endings. Prefer deleting or consolidating repeated
argumentative work over generating synonymous wording.
6. Check boundary placement. Scope must remain aligned across the paper, but
full limitation prose need not be repeated after every result. Preserve
essential local qualifiers and concentrate fuller study-boundary language
in the abstract ending, Methods boundary, Discussion/Limitations, and
conclusion when appropriate.
7. Present a numbered review table with location, function, issue type, why it
matters, and a proposed edit.
8. Wait for user approval before editing chapter files.
9. Apply only approved edits, then re-run the audit.
## Constraints
1. Never auto-fix chapter logic without user approval.
2. Preserve the author's argument and section structure.
3. Keep edits local to the flagged paragraph unless the user approves a broader rewrite.
4. Do not use a logic or style pass to repair an unstable research question,
evidence chain, application purpose, or contribution order.
5. No emoji.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!