Greptile

Canonical Agent Skills library for Claude Code, Cursor, Codex, and other agents.

Install

npx -y skills add lgtm-hq/ai-skills --skill greptile

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Run Greptile CLI for pre-push AI branch review. Use when asked for Greptile review, raycast/extensions pre-PR checks, or as part of the default dual pre-push workflow with coderabbit. Max 2 runs per change set.

SKILL.md

3.3 KB, as published. Nobody here has run it

Greptile CLI Review

Run Greptile (greptile) for automated branch review before push. Pair with coderabbit for the default pre-CI workflow — both tools catch different issues and mirror CI.

Relationship to analyze-* skills

The analyze-code, analyze-project, and analyze-tests skills are for manual or agent-driven pre-review checks—they run concrete commands (lintro, ripgrep, coverage, etc.) and produce structured findings before you open a PR.

This skill uses Greptile CLI for external AI branch review. Follow the lint and test skills (or the commit skill, which includes them) before invoking Greptile. Do not duplicate full analyze rubrics here.

Relationship to coderabbit

Default pre-push flow when CI runs both tools:

commit → [greptile ‖ coderabbit] → pr

Each step names the skill to follow.

Run greptile and coderabbit in parallel (‖) when possible. Greptile is typically faster (~1 min). Do not re-run on unchanged code.

Prerequisites

greptile login
greptile whoami

Install: npm i -g greptile or see Greptile CLI.

Commands

greptile review -b main --agent — Review current branch vs main (default for agents)
greptile review --diff — Findings inline with relevant code
greptile review --json — Machine-readable output (use instead of --agent for parsing)
greptile review --resume — Continue the latest unfinished review
greptile review show [ID] — Reopen a past review

--agent is an alias for --text (plain text, not JSON). Use --json when a tool needs structured output.

Rules

Max 2 Greptile runs per change set (workflow discipline)
Commit before running — Greptile reviews branch commits vs base; uncommitted changes are ignored
Default to -b main unless the repo uses a different default branch
Run from the repository root on the feature branch

Usage

When asked to run Greptile or as part of pre-push review:

Ensure changes are committed and lint/tests pass
Run greptile review -b main --agent (can run parallel with CodeRabbit)
Analyze findings; address critical/major issues
After fixes to Greptile-sensitive areas, optional verification pass
Track run count (max 2 per change set)

For Raycast extension PRs, run static greps from pr-raycast before the CLI review.

Differences from CodeRabbit

	CodeRabbit	Greptile
Scope	`all` / `committed` / `uncommitted`	Branch commits vs base only
Agent output	`--agent` → JSON stream	`--agent` → plain text; `--json` for machine-readable
Base branch	`--base main`	`-b main`