From dev-skills
This skill should be used when the user asks to "evaluate the project", "run a deep eval", "review this feature", "evaluate the auth module", "find issues", or wants multi-angle project evaluation with long-running critic agents. Dispatches skeptical evaluator agents that iterate through evaluation angles to find issues across code, architecture, security, UX, product, and design dimensions.
npx claudepluginhub aeghnnsw/cc-toolkit --plugin dev-skillsThis skill uses the workspace's default tool permissions.
Dispatch long-running skeptical critic agents to evaluate a project from multiple directions. Inspired by the principle that separating creation from evaluation — and tuning evaluators for skepticism — produces far better quality assessment than self-review.
Implements structured self-debugging workflow for AI agent failures: capture errors, diagnose patterns like loops or context overflow, apply contained recoveries, and generate introspection reports.
Monitors deployed URLs for regressions in HTTP status, console errors, performance metrics, content, network, and APIs after deploys, merges, or upgrades.
Provides React and Next.js patterns for component composition, compound components, state management, data fetching, performance optimization, forms, routing, and accessible UIs.
Dispatch long-running skeptical critic agents to evaluate a project from multiple directions. Inspired by the principle that separating creation from evaluation — and tuning evaluators for skepticism — produces far better quality assessment than self-review.
| User intent | Mode |
|---|---|
| "evaluate auth module", "review the payment flow", "check this feature" | Focus |
| "evaluate the project", "run a deep eval", "explore the codebase" | Explore |
If ambiguous, ask the user to clarify: "Would you like to focus on a specific feature, or explore the entire project?"
Target a specific feature or module. Critics iterate until no new issues are found, then self-terminate.
${CLAUDE_PLUGIN_ROOT}/skills/project-eval/references/angles.md — scan the 8 direction clusters${CLAUDE_PLUGIN_ROOT}/skills/project-eval/references/focus-mode.md — follow the complete workflowEvaluate the entire project for a time-limited period (default 30 minutes). Critics explore diverse angles until the deadline.
${CLAUDE_PLUGIN_ROOT}/skills/project-eval/references/angles.md${CLAUDE_PLUGIN_ROOT}/skills/project-eval/references/explore-mode.md — follow the complete workflowWhen all critics complete, read ${CLAUDE_PLUGIN_ROOT}/skills/project-eval/references/finding-format.md for the report format, then:
docs/eval/findings-<timestamp>-critic-*.md (use this run's timestamp)docs/eval/findings-<YYYYMMDD-HHMMSS>.mdRead angles.md for full details and calibration examples. The 8 clusters:
These are a starting menu — critics explore freely within their assigned direction.