Bita71/Agent-Workflow-Kit

Bita71/Agent-Workflow-Kit

Command Claude Code

Audit the project's AI surface: rules, skills, agents, commands, adapters. Report duplicates, gaps, stale rules, and inconsistencies; propose fixes without applying them.

Bita71/Agent-Workflow-Kit

Command Claude Code

Apply a worktree's diff into the base checkout — only when the user explicitly asks to transfer changes.

Bita71/Agent-Workflow-Kit

Command Claude Code

Implement the user's task in code.

Bita71/Agent-Workflow-Kit

Command Claude Code

Recommend which model and effort to use for the given work (often a .plan.md). Do not write code or run anything — just the recommendation.

Bita71/Agent-Workflow-Kit

Command Claude Code

Orchestrate planning through the Codex CLI: close questions, have Codex return the plan, reconcile it, and write the artifact as the host caller.

Bita71/Agent-Workflow-Kit

Command Claude Code

Runtime dependency audit. Report findings; do not change dependencies without an explicit request.

Bita71/Agent-Workflow-Kit

Command Claude Code

Propose only durable, reusable improvements to persist in the repo after this session. Do not apply them — the user copies, edits, or rejects.

Bita71/Agent-Workflow-Kit

Command Claude Code

Design review of the current diff: architecture, boundaries, public API, UX, simplification, performance.

Bita71/Agent-Workflow-Kit

Command Claude Code

Orchestrate a full review of the current diff: Codex correctness (+ security by risk signals) in parallel with a strong-model design review, then deduplicate into one report. No code is edited in this command.

Bita71/Agent-Workflow-Kit

Command Claude Code

Drive the whole task through the full pipeline as the orchestrator.

Bita71/Agent-Workflow-Kit

Command

Codex builds an implementation plan for the user's task and saves it as a Markdown artifact.