Review source code changes and report high-confidence bugs as severity-tagged notes.
npx skills add https://github.com/radareorg/radare2-skills --skill aireview-radareorg
Please help me install this Agent Skill. Skill: aireview Source: https://github.com/radareorg/radare2-skills/tree/main/dev/aireview Command: npx skills add https://github.com/radareorg/radare2-skills --skill aireview-radareorg
It helps you review source code changes in commits, functions, or files and produce only high-confidence, actionable findings in a consistent format.
Use the aireview skill to review the changed files in the provided commit or pull request and produce high-confidence severity-tagged findings.
High-intent search queries and answers about installing and using this skill.
To review pull request diffs for security and correctness, the skill analyzes changed lines in commits or files to identify high-confidence bugs, security concerns, and API contract violations. It gathers relevant change context before assessing potential functional issues.
Automated static analysis on source code changes is performed by targeting only modified lines within branches or pull requests to ensure focused diff inspection. This approach limits comments to high-confidence findings and reports them as structured severity-tagged task notes.
Yes, you can use radare2 for code review on specific commits and functions. The skill evaluates targeted source code changes to produce high-confidence, actionable findings in a consistent format, focusing its checks strictly on the modified lines within those commits or functions.
Diff inspection handles false positives during code review by enforcing a high-confidence threshold before reporting any findings. It limits comments to only high-certainty issues and emits concise, constructive descriptions with severity tags to maintain low-noise output.
Automated code review output for identified bugs uses a structured format consisting of severity-tagged task notes. Each finding includes a concise, constructive issue description limited to a maximum of two sentences to ensure actionable and low-noise reporting.
nilleReview code diffs for correctness, security, and conventions.
LAF-USReview pull request diffs and report actionable bugs with remediation steps.
macaughDetects injection, memory safety and authorization flaws in multiple languages.
cloudvoyantValidate code compliance with style guides and best practices.
