# review-hog-authoring

review-hog-authoring — a guide skill for writing custom PostHog Review pieces: extra PR perspectives, your own blind-spot sweep, and the bars that decide what gets published or fixed.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Windsurf (Devin Desktop), Claude.ai · 187 installs · ★ 101 repo stars
- Author: PostHog
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/PostHog/ai-plugin/tree/HEAD/skills/review-hog-authoring/SKILL.md
- Updated: 2026-10-24
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/review-hog-authoring

## What it does

PostHog Review checks pull requests on its own, and this skill is the manual for handing it extra reviewers. A run splits the PR into chunks, applies every enabled perspective to each chunk, sweeps for blind spots once, then keeps only findings that clear the validation bar. After publishing, the resolution stage works through open threads and commits changes that are both worth making and safe.

The skill walks the agent through the job: list the team's `review-hog-*` skills, read the canonical for the kind being written, settle the direction with the user, then create the `LLMSkill` row with `posthog:skill-create` rather than handing over text to paste. Names follow a fixed prefix per kind, such as `review-hog-perspective-<slug>` or `review-hog-validation-<slug>`, and that prefix is what the Code review tab and review runs look for. Later edits go through `posthog:skill-update`; duplicating a canonical with `skill-duplicate` drags seeded metadata along.

None of this lands without the PostHog MCP skill tools available to the agent. A newly written skill also starts switched off for its author: perspectives get toggled in Inbox → Code review → Perspectives, while a blind-spot, validation or resolution pick replaces whichever one is currently active. Versions are pinned when a run begins, so an edit mid-review shows up next time.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add PostHog/ai-plugin --skill review-hog-authoring -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/review-hog-authoring/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/review-hog-authoring/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add PostHog/ai-plugin --skill review-hog-authoring -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/review-hog-authoring/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add PostHog/ai-plugin --skill review-hog-authoring -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/review-hog-authoring/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install PostHog/ai-plugin skills/review-hog-authoring
```

2. **Manually: put the files in place**
- File: `.github/skills/review-hog-authoring/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add PostHog/ai-plugin --skill review-hog-authoring -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/PostHog/ai-plugin --path skills/review-hog-authoring
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/review-hog-authoring/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add PostHog/ai-plugin --skill review-hog-authoring -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/PostHog/ai-plugin/tree/HEAD/skills/review-hog-authoring/SKILL.md

## When to use it
- add a specialist lens to automated PR reviews
- tighten which findings get published on a pull request
- set your own bar for which review comments get implemented
- swap the final blind-spot sweep for one aimed at a known weak spot
- create a team `LLMSkill` from the agent instead of by hand

## Example prompts
- Create a custom review perspective for database migrations and call it `review-hog-perspective-migrations`.
- Show me the canonical blind-spot skill, then author our own sweep for the area we keep missing.
- Tune the validation criteria so only production-breaking findings reach the PR.

## Similar
- [ce-code-review](https://affpapa.org/en/nevibecode/skills/ce-code-review.md) — skill · 4100 installs
- [om-review-prs](https://affpapa.org/en/nevibecode/skills/om-review-prs.md) — skill · 3600 installs
- [review-pr](https://affpapa.org/en/nevibecode/skills/neolabhq-review-pr.md) — skill · 1300 installs
- [attach-review-to-pr](https://affpapa.org/en/nevibecode/skills/attach-review-to-pr.md) — skill · 1200 installs
- [github-pr-review](https://affpapa.org/en/nevibecode/skills/github-pr-review.md) — skill · 581 installs
- [review](https://affpapa.org/en/nevibecode/skills/parcadei-review.md) — skill · 523 installs

From the same set: [posthog/ai-plugin](https://affpapa.org/en/nevibecode/by/posthog/ai-plugin.md) — 82 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
