65 lines
2.2 KiB
Markdown
65 lines
2.2 KiB
Markdown
# Execute
|
|
|
|
Execute implementation plans with automatic agent selection, batch-level code review, and retrospective completion.
|
|
|
|
## Algorithmic Workflow
|
|
|
|
**Decision tree (follow exactly, no interpretation):**
|
|
|
|
1. Is this a plan execution request?
|
|
- YES → Continue to step 2
|
|
- NO → This command was invoked incorrectly
|
|
|
|
2. Does a plan exist to execute?
|
|
- YES → Continue to step 3
|
|
- NO → Run `/cipherpowers:plan` first to create implementation plan, then return here
|
|
|
|
3. **MANDATORY: Skill Activation**
|
|
|
|
**Load skill context:**
|
|
@${CLAUDE_PLUGIN_ROOT}skills/executing-plans/SKILL.md
|
|
|
|
**Step 1 - EVALUATE:** State YES/NO for skill activation:
|
|
- Skill: "cipherpowers:executing-plans"
|
|
- Applies to this task: YES/NO (reason)
|
|
|
|
**Step 2 - ACTIVATE:** If YES, use Skill tool NOW:
|
|
```
|
|
Skill(skill: "cipherpowers:executing-plans")
|
|
```
|
|
|
|
⚠️ Do NOT proceed without completing skill evaluation and activation.
|
|
|
|
4. **FOLLOW THE SKILL EXACTLY:**
|
|
- The skill defines the complete execution methodology
|
|
- Automatic agent selection (hybrid keyword/LLM analysis)
|
|
- Batch execution (3 tasks per batch)
|
|
- Code review after each batch
|
|
- Retrospective capture when complete
|
|
|
|
5. **STOP when execution is complete.**
|
|
|
|
## Why Algorithmic Workflow?
|
|
|
|
- **100% reliability**: No interpretation, no skipping plan creation
|
|
- **Skill integration**: Automatic discovery via Skill tool
|
|
- **Agent orchestration**: Skill handles agent selection and dispatch
|
|
- **Quality gates**: Code review checkpoints prevent cascading issues
|
|
|
|
## What the Skill Does
|
|
|
|
The executing-plans skill provides:
|
|
- Load and parse implementation plan
|
|
- Automatic agent selection (rust-agent, ultrathink-debugger, etc.)
|
|
- Batch execution with review checkpoints
|
|
- Code review after each batch (automatic dispatch to code-review-agent)
|
|
- Retrospective capture when work completes
|
|
- Integration with selecting-agents skill
|
|
|
|
**References:**
|
|
- Skill: `${CLAUDE_PLUGIN_ROOT}skills/executing-plans/SKILL.md`
|
|
- Agent Selection: `${CLAUDE_PLUGIN_ROOT}skills/selecting-agents/SKILL.md`
|
|
- Code Review: Automatic dispatch to cipherpowers:code-review-agent
|
|
- Integration: Seamless workflow → `/cipherpowers:brainstorm` → `/cipherpowers:plan` → `/cipherpowers:execute`
|
|
|