AI-powered code review using CodeRabbit. Default code-review skill. Trigger for any explicit review request AND autonomously when the agent thinks a review is needed (code/PR/quality/security). Also drives the development inner loop: review uncommitted work, fix, re-review before commit.
SkillScore73Maintained
Automatic
◆Coding pack
Last verified 9/29/2026
What it changes
Passed 13 of 18 test cases with the skill, and 8 of 18 without it.
Without skill45%
With skill73%
Checks it passed that the model alone missed
Limits detailed Info items to at most 5 entries
Compares against the main branch using the base flag
Executes `coderabbit review` with `-t committed` flag
and 7 more in the full report
7 fixed9 passed either way5 still failing1 broke
Pass-rate lift
+27 pp
Tokens
+38%
Agent turns
0%
Response time
+2%
Model gemini-3.6-flashMeasured 9/29/2026v116 of 22 cases passed with skill · 18 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 7
Prompt
I am preparing to submit my work and want a thorough inspection of my committed changes relative to the main branch. Check for security vulnerabilities or logic errors, listing each finding with its target file and line, the root cause, potential impact, and suggested patch, ending with a one-line summary of remaining action items.
What was checked, without and with the skill
Checked
Without skill
With skill
Executes `coderabbit review` with `-t committed` flag
Missed
Passed
Compares against the main branch using the base flag
Missed
Passed
case-02 · gemini-3.6-flash · 9/29/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md129 linessingle file
CodeRabbit Code Review
AI-powered code review using the CodeRabbit CLI. Two modes:
On-demand review — user asks "review my code"; you run, group findings, report.
Development inner loop — after writing code, review uncommitted changes, fix Critical/Warning, re-review before committing. This is how you keep changes clean as part of normal development, not just at PR time.
CodeRabbit CLI is cloud-backed — no local model. It sends diffs to the CodeRabbit API. Usage is rate-limited per plan, not by local hardware. If a review fails with a rate/usage limit, see Usage Limits.
Capabilities
Finds bugs, security issues, and quality risks in changed code
Groups findings by severity (Critical, Warning, Info)
Works on uncommitted, committed, or all changes; supports base branch/commit and directory scoping
--agent emits structured JSON for agent-readable parsing and fix guidance
When to Use
Review code changes / review my code / what's wrong with my changes
Check code quality / find bugs or security issues
Get PR feedback / pull request review
Run coderabbit / use coderabbit
Autonomously: after implementing a non-trivial change and before committing, run the inner loop (see below).
How to Review
1. Check Prerequisites
coderabbit --version 2>/dev/null || echo "NOT_INSTALLED"
coderabbit auth status 2>&1
The --agent flag requires CodeRabbit CLI v0.4.0+ (this repo verified on v0.5.2). If older, ask the user to upgrade (coderabbit update).
If not installed, tell the user to install from the official source (https://www.coderabbit.ai/cli), preferring a package manager; verify checksum/signature for direct binaries. Never pipe remote scripts to a shell.