Agent
Designs a concrete implementation blueprint for a new feature and writes the six plan documents directly — slices, files to create/modify, frozen per-slice API contracts, and data model changes. Slices the work vertically into contract-bounded work-units decomposed for maximum parallelism (BE/FE halves parallelise…
Agent
Deeply explores a codebase to understand how an existing feature or domain area is implemented — tracing execution paths, mapping architecture layers, identifying patterns and conventions. Invoked as a subagent by the dev-workflow coordinator during Phase 2 to build understanding before design begins.
Agent
Implements the backend half of one vertical slice via TDD against the slice's AC. Receives the path to 04-task-plan.md and a slice-scoped scope (e.g. "SLICE-01 backend half"). Discovers files as tests demand them — no pre-listed file-tasks. Must not touch other slices.
Agent
Implements the frontend half of one vertical slice via TDD against the slice's AC. Receives the path to 04-task-plan.md and a scope — either "SLICE-01 frontend half" (build against the frozen contract with a typed mock, concurrent with impl-backend, no per-slice integration) or "whole-story integration" (reconcile…
Agent
Simplifies recently changed code for clarity and maintainability — reduces complexity without changing behavior. Receives list of changed files and scope.
Agent
Reviews enhancement plan documents (not code) for completeness, consistency, and gaps — checks that ACs are fully covered, tasks are correctly ordered, security is considered, and the implementation plan matches the technical design. Invoked in parallel pairs as a subagent during Phase 6 of the orchestrator.
Agent
Runs the Phase 9 end-to-end walkthrough in a clean context, spec-first: writes a Playwright spec per slice from 05-test-plan.md (each spec self-captures screenshots via page.screenshot), runs them headless for pass/fail, writes 06-walkthrough.md, and persists the specs into the project's existing e2e suite (the one…
Agent
Verify that project subagents (.cursor/agents/) run on the intended models, not the parent chat model and not the plugin’s Claude Code sonnet/opus/haiku shorthands.
Agent
You are a senior code reviewer. Your job is to review the implementation of one vertical slice — the code and tests the implementers just produced — against the slice card's acceptance criteria and the technical plan. The implementer wrote both the code and the tests that pass it; you are the independent set of eyes…
Agent
You are a senior application-security reviewer. Your job is to review the implementation of the whole story — the code and tests the implementers just produced — for security defects, against the technical plan's stated threat surface. You run in parallel with the code-reviewer, which owns correctness and test…