agentic_development

Use when executing a slice plan wave by wave.

Use when validating generated slices.

Use after PR creation when requesting Codex or Copilot review.

Author acceptance tests from requirements before implementation.

Evaluate plans/tasks/outputs for schema and quality violations.

Audit API/schema/route/CLI contract changes for compatibility and migration needs.

Stage changes, commit, push, and create/update PR with review-ready hygiene.

Validate dependency additions/upgrades for risk and security.

Detect and remove AI slop patterns from code, docs, PRs, and comments.

Execute an epic plan wave-by-wave with dispatch, review, and merge checks.

Merge one wave's task branches one-by-one with verification.

Improve API, route, schema, event, CLI, and interface coherence.

Remove unused or unreachable code with safe reference checks.

Improve measurable performance with evidence, without speculative rewrites.

Perform behavior-preserving simplification and cleanup after green tests.

Generate release and migration notes after completion.

Generate a repo context map for later planning and implementation safety.

Request and track Codex/Copilot review comments on open PRs.

Convert feature context into epics, tasks, waves, and deterministic task artifacts.

Implement one generated task in strict worktree isolation and TDD style.

Review one task implementation against its spec with severity and pass/fail outcome.

Validate tasks are PR-sized, testable, independent, and reviewable.

Capture pre-implementation test status and environment assumptions.

Validate wave parallel safety and dependency ordering across task plans.

Use when the user asks Codex to understand a repository deeply before planning review/refactor work.

Use before commit or PR updates to remove AI-looking slop and scope creep.

Use to evaluate generated feature models, slice plans, wave plans, and run outputs.

Use after codebase-deep-analyzer or when the user asks for a high-level model of all repository features.

Use after completed waves or merged PRs.

Use when there is a feature model and the user wants reviewable/refactorable slices.

Use when a slice PR exists and CI must be inspected or repaired before an optional explicit merge.

Use after a slice branch has verified changes.

Use after slice-review-workflow, when applying valid findings and simplifying code.

Use when reviewing one slice.

Use after slice generation to determine parallel waves.