boardclaude

Run a multi-agent panel audit on the current project (or a specified target).

Compare two strategy branches (or two specific audits) side by side.

Implement fixes from audit action items using parallel batch execution

Create a git worktree for a parallel strategy experiment. This allows you to

Initialize BoardClaude for the current project. Creates the directory structure,

Merge a strategy branch into the current branch (typically main). Archives

Run a lightweight, single-agent review using one specific agent persona.

Documentation, developer relations, and accessibility evaluator inspired by Ado Kukic's role as DevRel Lead at Anthropic. Assesses README quality, example coverage, onboarding experience, accessibility compliance, and community readiness. Use when running hackathon panel audits.

Architecture and verification specialist inspired by Boris Cherny's public engineering philosophy. Evaluates project architecture, verification rigor, compound learning patterns, and simplicity. Use when running hackathon panel audits.

Product and user impact evaluator inspired by Cat Wu's role as Product Manager for Claude Code. Assesses user value, adoption path, narrative coherence, and market fit. Use when running hackathon panel audits.

Focused implementation agent that pulls action items from the shared TaskList. Claims tasks, implements fixes, runs quick validation, reports results via task metadata and SendMessage. Loops until no work remains.

Community impact, narrative quality, and integration evaluator inspired by Jason's public role at Anthropic. Assesses whether the project is accessible, well-communicated, internationally aware, and production-ready. Use when running hackathon panel audits.

Frontend, developer experience, and code quality evaluator inspired by Lydia Hallie's public expertise in React patterns, TypeScript, and web performance from patterns.dev and Frontend Masters. Evaluates code quality, DX, performance, and modern patterns. Use when running hackathon panel audits.

Panel moderator and report generator. Merges findings from all panel agents into a unified evaluation report with composite scores, radar chart data, agreements, disagreements, and prioritized action items. Use after all panel agents have completed their evaluations.

AI innovation and model capability evaluator inspired by Thariq Shihipar's background in AI systems and novel applications. Assesses how deeply and creatively the project uses Opus 4.6 features, whether it demonstrates emergent behavior, and efficiency of AI usage. Use when running hackathon panel audits.

Run Playwright-based accessibility, visual, and performance audits on the dashboard. Produces Lighthouse scores and axe-core results. Triggers on "accessibility", "a11y", "lighthouse", "visual test", "screenshot".

Run multi-agent panel audits on codebases. Use when evaluating code quality, running board reviews, or scoring projects against configurable criteria. Triggers on "audit", "evaluate", "review", "score", "panel", "board".

Full orchestrator for the fix pipeline. Loads audit data, filters action items, runs validation baseline, creates native tasks with dependencies, spawns a fix-worker pool that self-organizes via TaskList, validates batches through gate checkpoints, retries failures, updates the ledger, and reports results. Triggers on "fix", "implement", "resolve", "address findings".

Guide for creating new evaluator agent personas. Use when building custom panels, adding agents to existing panels, or refining existing personas. Triggers on "create persona", "new agent", "build evaluator", "add judge", "persona", "new panel member".

Manage branching timeline for BoardClaude audits. Handle forks, merges, comparisons, and timeline visualization data. Triggers on "fork", "branch", "compare", "merge", "timeline", "strategy", "worktree".

Run project validation suite (types, tests, lint, format) and produce structured results for agent evaluation context. Triggers on "validate", "check", "verify", "test", "lint".

2 hooks across 2 events