review-hog-perspective-logic-correctness — Claude Code skill
review-hog-perspective-logic-correctness reviews one slice of a pull request for correctness: arithmetic, branches, data mapping, database queries. Useful when you want a focused second pass before merging.
no license 187 installs ★ 101 updated 2026-10-24
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a claude-codeWhat review-hog-perspective-logic-correctness does
The skill reads a chunk of a pull request and hunts for mistakes in the logic: wrong arithmetic, branches that miss a case, fields dropped during mapping, queries that come back with the wrong rows. State updates in frontend code get a look too, as do blocking SQL calls issued from async code. Prompts for an LLM are in scope as well, including how the answer is parsed and how token limits are handled.
What you get is a checklist over four areas plus a handful of rg searches: for calculation code, for conditionals and switch, for SQL keywords, for state setters. Findings land only on files that carry real logic. Tests and config are read for context and never flagged.
It is one of several viewpoints running over the same diff in parallel. Security and performance belong to other perspectives, and a later dedup pass sorts out whatever overlaps. Run on its own, it covers just one part of a review.
What it helps with
- Catch an off-by-one error before merge
- Check that a JOIN and its WHERE clause return the right rows
- Spot a field that disappears during mapping between layers
- Trace an unexpected state mutation in a component
- Confirm a sync SQL call is not running inside async code
Example prompts for the agent
- Review the changes in
billing/calc.pythrough the logic and correctness perspective - Check this diff for missed edge cases and broken queries
- Run a correctness pass over
api/handlers/orders.ts
How to install review-hog-perspective-logic-correctness
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a claude-code2Manually: put the files in place
~/.claude/skills/review-hog-perspective-logic-correctness/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/review-hog-perspective-logic-correctness/
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a codex2Manually: put the files in place
.agents/skills/review-hog-perspective-logic-correctness/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a cursor2Manually: put the files in place
.cursor/skills/review-hog-perspective-logic-correctness/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install PostHog/ai-plugin skills/review-hog-perspective-logic-correctness2Manually: put the files in place
.github/skills/review-hog-perspective-logic-correctness/SKILL.md1Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a gemini-cli2Platform CLI
gemini skills install https://github.com/PostHog/ai-plugin --path skills/review-hog-perspective-logic-correctnessManually: put the files in place
.devin/skills/review-hog-perspective-logic-correctness/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about review-hog-perspective-logic-correctness
How do I install review-hog-perspective-logic-correctness in Claude Code?
Run npx skills add PostHog/ai-plugin --skill review-hog-perspective-logic-correctness -a claude-code. The skill file lands at ~/.claude/skills/review-hog-perspective-logic-correctness/SKILL.md.
What does review-hog-perspective-logic-correctness do?
It goes over a pull request chunk looking for correctness bugs: faulty arithmetic, unhandled boundaries, bad data mapping, query mistakes, state bugs. It reports logic issues and nothing else.
Does it cover security or performance?
No. Vulnerabilities and API contracts belong to another perspective, and performance plus error handling to a third. Anything two perspectives both report gets merged by a separate dedup step.