菜单
Transition from a hands-on "bricklayer" to a high-level "architect" by managing a fleet of autonomous AI agents. Use this when you need to scale engineering output with a small team, handle repetitive migrations/bug fixes, or onboard engineers to complex legacy codebases.
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
MUST BE USED when analyzing meeting transcripts or notes. Classifies input type, attributes speakers, and verifies names against colleagues.json. Works with routing-brains and summarizing-meetings skills for complete workflow.
A framework for evaluating whether a product's success is mathematically possible based on its cost structure, acquisition assumptions, and unit economics. Use this when vetting a new job opportunity, evaluating a startup's viability, or determining if a failing product can be "iterated" to success.
A structured method for gathering high-signal, contextual feedback from a trusted network to validate ideas or navigate complex career decisions. Use this when facing a "fork in the road" career choice, prioritizing product features, or narrowing down topics for public speaking or content.
Create and maintain individual colleague communication profiles in Markdown. Use when the user wants to document a person's communication style, tone preferences, decision style, sensitivities, channels, response cadence, or examples of good vs bad communication, and store it as a per-person profile under input/org/profiles/.
Gather context through high-leverage eigenquestions when information is missing or ambiguous. ACTIVATE THIS SKILL when: (1) context is missing or unclear, (2) user asks something requiring unstated assumptions, (3) working within a brain without recent context, (4) starting new work or strategic decisions, (5) multiple valid approaches exist without clear direction, (6) user says "clarify", "help me think through", or "what should I consider". This skill asks ONE discriminating question at a time to unlock maximum downstream clarity.
Drafts concise, professional PM emails using the CRAFT method. Use when user needs to compose, draft, write, or rewrite professional email communication including follow-ups, status updates, escalations, and stakeholder messages.
| name | agentic-engineering-workflow |
| description | Transition from a hands-on "bricklayer" to a high-level "architect" by managing a fleet of autonomous AI agents. Use this when you need to scale engineering output with a small team, handle repetitive migrations/bug fixes, or onboard engineers to complex legacy codebases. |
This workflow enables you to transition from manual implementation to high-level system architecture by managing autonomous AI agents (like Devin) as "junior buddies." By shifting implementation to agents, you can scale a small team (e.g., 15 engineers) to handle the output of a much larger organization, aiming for 25% to 50% of pull requests to be AI-generated.
Most engineering time is spent on "bricklaying": debugging Kubernetes errors, fixing port issues, or writing boilerplate code. Your goal is to move to "architecting": defining the problem precisely, mapping out the solution, and specifying trade-offs, while the agent handles the execution.
Do not hand agents "problems" (ambiguous high-level goals); hand them "tasks" (well-defined, verifiable units of work).
Do not watch the AI work action-by-action. To achieve massive productivity gains, you must manage multiple agents in parallel.
Integrate the agent into your existing human workflows to maintain quality and oversight.
Use agents to bridge the knowledge gap for human engineers.
Example 1: Bug Reproduction and Fix
Example 2: Feature Implementation
