Coordinates parallel specialist audits of code diffs into a unified report.
npx skills add https://github.com/hawkyre/hawk-skills --skill code-audit-hawkyre
Please help me install this Agent Skill. Skill: code-audit Source: https://github.com/hawkyre/hawk-skills/tree/main/skills/code-audit Command: npx skills add https://github.com/hawkyre/hawk-skills --skill code-audit-hawkyre
Auditing code diffs and changes across large repos is error-prone and time-consuming. This skill coordinates a fan-out of independent specialists to evaluate code on its own merits, ensuring broader coverage and objective assessment without knowledge of the user's goal or plan.
Run the code-audit orchestrator on a target diff to generate a consolidated report.
High-intent search queries and answers about installing and using this skill.
Automate code review for diffs by running parallel independent subagents that evaluate logic, security, and architecture simultaneously. This fan-out approach ensures broad coverage and objective assessment across large repos, merging all findings into a unified structured report.
Parallel code audit is a fan-out orchestration where independent specialist subagents inspect code diffs across logic, security, simplification, research, and architecture. It applies goal-blind, context-free evaluation to prevent bias and merges findings into a cohesive audit report highlighting FIXes, NOTES, and QUESTIONS.
Yes, you can audit code changes across different programming languages and architectures. The orchestration applies goal-blind evaluation to code diffs and PRs regardless of language, ensuring consistent quality and objective assessment across varied software engineering stacks.
Generate a structured audit report from code diffs by running the code-audit orchestrator on a target diff. It coordinates parallel specialist audits and merges findings into a cohesive report that highlights FIXes, NOTES, and QUESTIONS for human review.
The best way to prevent bias during security and architecture code reviews is using goal-blind, context-free evaluation. By running independent specialist subagents that evaluate code strictly on its own merits without knowledge of the user's goal, you ensure objective and consistent assessment.
ben-mad-jlpDebug bugfix items by writing failing tests and coordinating parallel subagents.
tomzxAudit GitHub issues for completeness, clarity, and acceptance criteria quality.
Mo7amed-MahmoudRun parallel dependency, static, and privacy scans and emit severity findings.
ivanshtokovCoordinate multiple agents for multi-domain analysis with parallel and sequential patterns.
