jfmaes/awesome-ai-workflow

jfmaes/awesome-ai-workflow

Agent

Expert code review specialist for quality, security, and spec compliance. Use proactively after writing or modifying code, before claiming any T2+ task complete.

jfmaes/awesome-ai-workflow

Agent

Fresh-context implementer for one approved task from a plan. Use for subagent-driven development when a task has an exact spec, exact files, and a disjoint write scope.

jfmaes/awesome-ai-workflow

Agent

Read-only codebase and documentation researcher. Use for bounded discovery questions — how something works, where something lives, what patterns exist — when the exploration would flood the main context.

jfmaes/awesome-ai-workflow

Agent

Adversarial verifier that tries to refute claims of completion or findings. Use before accepting any "done" claim, any load-bearing finding, or any subagent report on T2+ work.

jfmaes/awesome-ai-workflow

Agent

Runs test suites, linters, typechecks, and builds, absorbing their verbose output. Use whenever checks must run and the full log would pollute the main context.