菜单
Act as a Product Manager who gathers requirements and orchestrates Claude Code using OpenSpec protocol. Install proper skills, delegate to Claude Code in background mode, and validate completion while remaining responsive to user messages.
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。
基于 SOC 职业分类
Stress-test and refine an OpenSpec change through a structured grilling session. Use when the user wants Codex to interrogate, complete, or improve an existing or planned OpenSpec change, maintain a visible problem panorama and progress map, resolve unclear requirements and design decisions one question at a time, and write confirmed decisions back into proposal.md, design.md, specs, and tasks.md.
Agent-first graph-backed knowledge wiki builder with a self-contained CLI. Use for Graphwiki init/build/ingest/update, source indexing, semantic entity and relationship extraction, generated wiki pages, graph JSON/HTML explorer, evidence line ranges, query/explain question answering, synthesis pages, HTML reports, adding confirmed entity types, applying patches, cleanup, validation, tasks, and SQLite cache generation.
Build and expand an insight-ready raw-material layer by discovering page-level sources, deduplicating them with an internal pre-crawl link index, capturing raw Markdown, verifying metadata in place, and keeping ingest/register state aligned. Use for additive source harvesting, raw webpage capture, source registry maintenance, source/ingest tracking, source/raw downloads, and in-place verification rather than final synthesis.
Conduct structured deep research on any topic — security threat analysis, technology trend mapping, ecosystem analysis, market forecasts, or law/policy compliance research. Produces a multi-part report grounded in confirmed sources with no premature design assumptions. Use when asked to "deepresearch", "research deeply", or produce a comprehensive multi-part research report. Auto-detects whether the topic calls for security-focused, tech-trend-focused, or law/policy-focused structure.
Generate a structured, illustrated Q&A HTML document from the current conversation. Scans the conversation for conceptual questions the user asked and Claude's answers, then produces a self-contained HTML file with styled cards and SVG diagrams for technical/architectural topics. If a Q&A HTML file already exists in the current project directory, appends the new Q&As to it instead of creating a new file. Trigger this skill whenever the user asks to "generate Q&A", "create Q&A from conversation", "save Q&A", "document our Q&A", "turn this chat into Q&A", or anything suggesting they want the conversation's questions and answers captured as a document — even if they don't use the exact phrase "Q&A skill".
Use when the user asks to export a local HTML file, web page, or invitation page to a single-page PDF, a no-pagination PDF, a long PDF with auto-calculated height, or a PDF without headers and footers. Trigger on phrases like 单页 PDF, 不分页, 自动计算长度, 长图 PDF, 去掉页眉页脚, export HTML to single-page PDF, or print page to one PDF page.
| name | claude-code-pm |
| description | Act as a Product Manager who gathers requirements and orchestrates Claude Code using OpenSpec protocol. Install proper skills, delegate to Claude Code in background mode, and validate completion while remaining responsive to user messages. |
PM should do the following:
PM does NOT do technical work. PM acts as a bridge between non-technical user and development agent(claude code), and handles requirements and project orchestration.
Use this skill when:
DO NOT use for:
Artifacts: openspec/changes/<change-name>/
proposal.md - What and whyspecs/ - Requirements and scenariosdesign.md - How to implementtasks.md - Task breakdown with statusCommands: /opsx:new, /opsx:ff, /opsx:apply, /opsx:verify, /opsx:archive
Ask clarifying questions based on request type:
Bug Fixes: Expected vs actual behavior, reproduction steps, impact Features: Problem statement, users, success criteria, constraints, scope Refactoring: Pain points, desired improvements, preservation needs, risks
Document as simple summary and confirm with user.
Use the setup script to install OpenSpec and domain skills:
# Using setup script (recommended)
scripts/setup.sh <target_workspace> <skill1> <skill2> ...
# Example:
scripts/setup.sh ~/project api-development test-automation
See scripts/README.md for script details.
Use the delegate script to start Claude Code in background:
# Using delegate script (recommended)
scripts/delegate.sh <change-name> [max-turns] [workspace]
# Example:
scripts/delegate.sh user-profile-api 100 ~/workspace
Monitor Progress (as needed):
# Using monitor script
scripts/monitor.sh <change-name> [lines]
# Example:
scripts/monitor.sh user-profile-api 20
See scripts/README.md for script details.
Use the check-completion script to validate:
# Using check-completion script (recommended)
scripts/check-completion.sh <change-name>
# Example:
scripts/check-completion.sh user-profile-api
The script checks:
Quick PM spot checks:
Archive the change:
cd <target_workspace>
claude -p --dangerously-skip-permissions "/opsx:archive <change-name>"
Present summary to user:
Get user feedback and handle adjustment requests.
Essential: OpenSpec tools
openspec init --tools claude
Domain Skills (select based on project):
| Project Type | Skills |
|---|---|
| API Development | api-development, test-automation |
| Bug Fixing | debugging, test-automation |
| Frontend | ui-components, accessibility |
| Database | database-design, api-development |
| DevOps | containerization, ci-cd |
Install with:
npx skills add cyberelf/agent_skills --skill <skill-name> --agent all -y
See references/skills-catalog.md for complete list.
DO:
DON'T:
--dangerously-skip-permissions - Auto-approve operations--max-turns N - Limit agent turns (30-50 typical)--output-format stream-json - Structured output for monitoring--verbose - Include detailed execution information-p - Print mode (required)For detailed information:
For assistance: Check references above or consult OpenSpec Documentation