Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Fix try-catch anti-pattern in Output SDK workflows. Use when retries aren't working, errors are being swallowed, seeing unexpected FatalError wrapping, or when step failures don't trigger retry policies.
Fix Zod schema import issues in Output SDK workflows. Use when seeing "incompatible schema" errors, type errors at step boundaries, schema validation failures, or when schemas don't match between steps.
Audit an existing eval suite for trustworthiness. Use when inheriting evals, suspecting evals miss real failures, or after significant pipeline changes.
Systematically review workflow traces to identify failure modes before building evaluators. Use when starting an eval project, after significant pipeline changes, or when production quality drops.
Post-flight validation for Output SDK workflow operations. Systematic verification of step completion, convention compliance, quality validation, and deliverable verification.
Analyze Output SDK workflow execution traces. Use when debugging a specific workflow, examining step failures, analyzing input/output data, understanding execution flow, or when you have a workflow ID to investigate.
Use when hitting a specific PyPhi error or surprising behavior — floating-point phi comparisons, pyphi_config.yml not being picked up, missing optional-dependency import errors, TPM format confusion, system state validation failures, cache memory growth, parallel computation not engaging, or mixing IIT 3.0/4.0 code paths.
Interactive git diff annotation review. Generates a cleaned-up diff, opens in editor for user annotations, and addresses feedback in a loop. Activates on "git review", "review changes", "review my changes", "annotate changes", "interactive review".
Systematic root cause analysis for errors, bugs, and unexpected behaviors using 5-Why methodology. Use when user reports errors, build failures, test failures, performance issues, integration problems, or any "it's not working" scenarios.
Plan and run a design-QA / "taste test" of llamactl changes against a real backend. Cooperatively builds a small matrix of cases worth eyeballing, runs them against a chosen backend (test environment by default, local kind+tilt for new API contract changes), and writes a design-review report to `thoughts/shared/qa/`. Use this for changes to `llamactl` commands, output formats, auth, or control-plane API contracts. For UI/template smoke tests, use `llamactl_browser_test` instead.