- Home
- Skills
- Code Quality & Review
- Risk-Tiered Code Review
Risk-Tiered Code Review
Review pull requests and diffs by blast radius, verify the evidence, and report separate merge and launch readiness verdicts.
Free
Works with the AI Agents you already use
Risk-Tiered Code Review
Example session with this skill installed
Review the staged changes against HEAD. Check whether the new price filter handles zero, classify the risk, and tell me if it’s ready to merge.
- Read your context and instructions
- Compiled the risk-tiered code review
Illustrative example using a fictional repository.
Scope: Staged changes against HEAD 8ac91de; only src/search/filters.ts and its tests are in scope.
Risk: R1 (25 points) for an isolated runtime change.
Merge: BLOCKED. Launch: NOT ASSESSED.
Finding
[P2] Preserve the zero-price filter — src/search/filters.ts:42. The new truthiness check skips a valid maxPrice=0. The search then includes paid items instead of returning only free items. Check whether the value is undefined before omitting the filter, and add a regression test for zero.
Risk and evidence
| Factor | Score | Reason |
|---|---|---|
| Surface | 5 | One search filter |
| Data and privilege | 10 | Changes handling of a non-sensitive query value |
| Reversibility | 0 | The filter has no persistent effects |
| Contracts | 0 | No interface change |
| Exposure | 10 | The changed branch is enabled for all searches |
The staged tests cover a positive maximum and an omitted maximum. They do not exercise zero. Test execution and independent review were not verified for this snapshot.
Next steps
Fix the condition, run the search tests with the new zero-price case, and obtain an independent review of the updated diff.
Connects securely to your tools. The creator never sees your data.
About this skill
Review a pull request, commit range, or local diff by asking three questions: What could this change affect? What did the checks actually verify? What still needs to happen before merge or launch?
- Pin the review to an exact revision or local snapshot so findings and test results refer to the same code.
- Score the change across five factors: surface, data and privilege, reversibility, contracts, and exposure. Use the resulting tier to decide how deeply to review and who needs to sign off.
- Trace affected callers and failure paths. Distinguish verified results from missing checks, stale runs, and unconfirmed claims.
- Report actionable findings with file and line references. Assess launch readiness separately when requested.
Provide a PR or diff and any acceptance criteria or repository rules. The review is read-only: it does not edit code, post reviews, approve, merge, or deploy.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.
Requires the full change and relevant evidence to verify readiness. An agent review cannot replace a required human approval. The skill does not edit code, post reviews, approve, merge, deploy, or probe live systems.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 4 days ago
- Free with an Agensi account
Permissions
Reads the requested diff, repository files, and available check results. The shell can inspect local Git state or run authorized checks. Remote PRs can use a connected provider. The skill does not edit code, post reviews, merge, or deploy.