用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/az9713/pilot-in-command-workflow --skill pic-conflict命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | pic-conflict |
| description | Escalate a conflict between PICs to the user for resolution |
| args | [conflict summary] |
You are escalating a conflict to the user for resolution. Follow this protocol exactly.
Conflict summary: $ARGS
If no summary was provided, ask the user to describe the conflict.
This skill is used when:
Read .pic/state.json to verify a workflow is active.
If no workflow:
No active workflow. Conflicts can only be recorded within a workflow.
And stop.
Read existing conflicts in .pic/conflicts/ to determine the next ID.
Format: CON-[NNN] where NNN is zero-padded (e.g., CON-001)
Collect or ask for:
Write to .pic/conflicts/CON-[NNN].md:
# CON-[NNN]: [Brief Title]
**Date**: [ISO date]
**Workflow**: [workflow id]
**Phase**: [current phase]
**Status**: Open
## Summary
[Brief description of the conflict]
## Conflicting Positions
### Position A: [Name/Source]
**Advocate**: [Which PIC or entity]
**Position**: [What they recommend]
**Evidence**:
- [Supporting evidence 1]
- [Supporting evidence 2]
**Rationale**: [Why they believe this is correct]
### Position B: [Name/Source]
**Advocate**: [Which PIC or entity]
**Position**: [What they recommend]
**Evidence**:
- [Supporting evidence 1]
- [Supporting evidence 2]
**Rationale**: [Why they believe this is correct]
## Analysis
### Points of Agreement
- [What both sides agree on]
### Core Disagreement
[The fundamental issue causing conflict]
### Stakes
[What depends on this decision being made correctly]
## Attempted Resolution
[What was tried before escalation]
## Options for Resolution
### Option 1: Accept Position A
- **Impact**: [What happens if we choose this]
- **Risk**: [Potential downsides]
### Option 2: Accept Position B
- **Impact**: [What happens if we choose this]
- **Risk**: [Potential downsides]
### Option 3: Synthesis (if possible)
- **Approach**: [How to combine elements of both]
: [What we give up]
Please review the positions above and indicate your decision:
[ ] Accept Position A
[ ] Accept Position B
[ ] Accept Synthesis
[ ] Provide alternative direction
---
: [pending]
: [pending]
: [pending]
: [pending]
Add conflict reference to .pic/state.json conflicts array:
{
"id": "CON-[NNN]",
"summary": "[summary]",
"phase": "[current phase]",
"status": "open",
"timestamp": "[ISO]"
}
Append to .pic/status-log.jsonl:
{"timestamp": "[ISO]", "event": "conflict_escalated", "conflict": "CON-[NNN]", "summary": "[summary]"}
Display the conflict clearly:
## Conflict Escalation
A conflict requires your input to proceed.
**ID**: CON-[NNN]
**Phase**: [current phase]
### The Disagreement
[Clear statement of what's in conflict]
### Position A: [Name]
[Summary of position with key evidence]
### Position B: [Name]
[Summary of position with key evidence]
### What's at Stake
[Why this decision matters]
---
Please indicate how you'd like to proceed:
1. Accept Position A
2. Accept Position B
3. Propose a different approach
Your decision will be recorded and the workflow will continue accordingly.
When the user provides a decision:
resolvedWhile a conflict is open and unresolved:
/pic-statusPre-flight dependency risk assessment for the planning phase. Use BEFORE writing code when choosing libraries, designing architecture, or setting up a new project. Prevents dependency hell, version conflicts, and abandoned-library traps. Especially critical for ML/AI projects with heavy dependencies.
View detailed audit log for workflow - shows agent executions, tool usage, and decision trail
Record a formal decision document with rationale and alternatives