Run eight Devil's Advocate agents in parallel to critique plans and code diffs.
npx skills add https://github.com/greenheadHQ/nixos-config --skill run-da
Please help me install this Agent Skill. Skill: run-da Source: https://github.com/greenheadHQ/nixos-config/tree/main/modules/shared/programs/claude/files/skills/run-da Command: npx skills add https://github.com/greenheadHQ/nixos-config --skill run-da
Automates rigorous, multi-domain review of plans and code by running eight specialized Devil's Advocate agents in parallel to surface design flaws, quality issues, and security concerns.
Invoke the run-da skill with a mode selector (for_plan, for_pr, or both) and an optional fresh modifier.
High-intent search queries and answers about installing and using this skill.
Automated code review on a pull request is performed by invoking the run-da skill with the for_pr mode, which analyzes code diffs across eight specialized Devil's Advocate domains in parallel to surface concrete security and quality issues.
Validating a software design plan is done using the for_plan mode, which runs parallel Devil's Advocate agents to critique the plan across domains like YAGNI, NGMI, and CONSISTENCY, providing actionable mitigation steps to improve design robustness.
Yes, you can review both a project plan and code diff simultaneously by using the both mode selector, allowing the eight parallel review agents to evaluate planning logic and implementation quality concurrently.
To avoid context carryover between reviews, you can apply the optional fresh context modifier, which ensures the agents evaluate the current plan or pull request independently without prior conversation history.
An AI-assisted pull request review checks eight specific quality assurance domains: YAGNI, NGMI, HALLUCINATION, SECURITY, SIDE_EFFECT, CONSISTENCY, READABILITY, and CLEAN_CODE.
Code review findings are reported with concrete file:line references and actionable mitigation steps, logging all issues per the Devil's Advocate protocol to ensure developers can directly locate and resolve the identified design flaws.
nbbaierReview pull requests with parallel AI agents and Git worktrees.
dogearedGenerate multiple AI-driven UI design variants and display them side-by-side in a browser for rapid evaluation.
plurigridEnforce tests, documentation, deadlines, and quality gates before commits and shipping.
RangeyRoverRun static analysis across multiple languages and generate prioritized issue reports.
