| name | ai-native-engineering-org |
| description | Practices for adapting engineering planning, context sharing, code review, and team operating norms once agentic coding becomes the default. |
Instructions
Use this skill to audit and redesign engineering workflows when agentic coding dramatically increases code output and shifts bottlenecks toward verification, review, and safety.
- Identify the workflow you want to change
- Choose the “noisiest” workflow (high friction, slow, or frequently discussed).
- Evaluate whether it still serves its purpose
- Be explicit about the purpose and what signals success.
- Decide: keep, automate, or drop
- If it still matters, ask whether the step can be automated (partially or fully).
- If it no longer matters, remove it.
- Apply the post’s operating changes
- Planning: move from long-range roadmaps to just-in-time planning; prefer PR discussions and prototypes.
- Context gathering: ask Claude first; only then ask humans if needed; consider automation.
- Code review: let Claude handle style/lint, tests, and obvious bugs; reserve human review for trust boundaries, security-sensitive code, legal risk, and product taste.
- Team structure: expect role boundaries to blur; keep teams flat; dogfood the product; kill processes that stop working.
- Roll out new norms
- Define a small set of non-negotiable principles.
- Let pods adapt the details for their local workflows.
- Track whether changes stick
- Track onboarding ramp time, PR cycle time, and whether commits are Claude-assisted.
Examples
