Plan-first development partner for Claude Code: interview-driven plans, tier-routed agents (haiku/sonnet/opus), and an adversarial review chain that verifies through real usage.
Agent
Deep, parallel-friendly codebase research specialist. Use proactively when exploration needs more than three queries, covers multiple naming conventions, requires deep file traversal, or benefits from LSP/AST-grep precision. Triggers on questions like "where is X defined", "who calls Y", "find all usages of Z", "how…
Agent
External documentation and OSS research specialist. Use proactively when the question involves an unfamiliar library, framework, or external API; needs official documentation; references third-party code patterns; or asks about open-source implementations. Triggers on questions like "how do I use [library]?", "what's…
Agent
Strategic technical advisor for architecture decisions, deep debugging stalls, second-opinion review, security or performance concerns, and reuse-vs-build trade-offs. Use when the question requires elevated reasoning beyond the orchestrator's bandwidth: multi-system tradeoffs, unfamiliar patterns, 2+ failed fix…
Agent
6-stage post-implementation reviewer for plans of complex complexity. Reads the plan file path plus modified-files list, runs Stages 1-4 identical to ac:plan-code-review (compliance L1/L2/L3 + spec + quality + simplify) then Stage 5 cross-layer integration (integration trace across module boundaries + caller impact…
Agent
4-stage post-implementation reviewer for plans of standard complexity. Reads the plan file path plus the modified-files list passed in the prompt, verifies the implementation against plan claims. Stage 1 compliance (L1 Exists / L2 Substantive / L3 Wired + Must NOT Have + Scope Fidelity, gates everything), Stage 2 spec…
Agent
Adversarial independent reviewer for plans of complex complexity. Two-pass: Pass 1 runs the same four blocker checks as ac:plan-reviewer, Pass 2 stress-tests the plan along eight adversarial dimensions (deep reference verification, executability stress-test, cross-task dependency analysis, tier challenge, QA…
Agent
Independent second-eye reviewer for plans of standard complexity. Reads a .ac/plans/ /plan.md path as the sole prompt, verifies reference validity, executability, internal consistency, and tier fitness. Returns [OKAY] or [REJECT] with up to a step-scaled cap of blocking issues, plus an uncapped Non-blocking…
Agent
Plan step executor for junior-high tier steps: junior-shaped work that sits at the borderline of coupling or context depth, or that codebase state escalated. Same capabilities and same reporting contract as ac:plan-worker-junior, run at high effort instead of medium, so a planner facing borderline work has somewhere…
Agent
Standard plan step executor for junior tier steps. The default worker tier: 1-3 file changes, business logic, pattern application, refactor-with-pattern, framework idiom application. Reads broader context than quick (callers, tests, imports), applies the step Description in the codebase's style, runs full verification…
Agent
Mechanical plan step executor for quick tier steps. Single-file or near-single-file changes that do not require surrounding-code understanding. Config edits, renames, scaffold work, single-line fixes, doc-block additions. Reads exactly the files the briefing names, applies the precise change, runs the verification…
Agent
Senior plan step executor for senior tier steps. Cross-layer changes, architecture, migration, complex edge cases, self-verification needs. Receives high-level intent and architectural constraint from the briefing, designs the solution within those constraints. Reads broadly (multiple modules, callers, tests, type…