Community写作与编辑github.com

jpantsjoha/ai-native-developer-experience

Team-wide AI harness adoption plugin, \w operating model, onboarding and delivery standards coherent human-agent outcomes from day one.

ai-native-developer-experience 是什么?

ai-native-developer-experience is a Claude Code agent skill that team-wide AI harness adoption plugin, \w operating model, onboarding and delivery standards coherent human-agent outcomes from day one.

兼容平台~Claude Code~Codex CLI~Cursor
npx skills add jpantsjoha/ai-native-developer-experience

Installed? Explore more 写作与编辑 skills: steipete/notion, langchain-ai/langchain, bytedance/podcast-generation · View all 6 →

在你喜欢的 AI 中提问

打开一个已预加载此 Agent Skill 的新对话。

文档

Sitrep

Situation reports are for the person who was not in the room. Write for them.

A sitrep is not a diary of activity. It is a compressed, actionable picture of current state. Three questions: what is done, what is blocked, what is next. Anything that does not answer one of those three questions does not belong in the sitrep.

When to use

  • Daily standup synthesis from work logs, commits, or ticket state
  • End-of-sprint summary for stakeholders
  • Incident situation report during or after an active incident
  • Handoff summary when switching context or handing work to another agent or person
  • When a stakeholder asks "where are we?" and the honest answer is "it's complicated"

Procedure

  1. Gather the current work state — read from available sources:

    • Open and recently closed issues / tickets
    • Recent commits (last 24-48 hours for a standup; last sprint for a sprint summary)
    • Project status and roadmap documents (e.g. docs/operating-model/PROJECT-OPERATING-PROFILE.md, docs/ROADMAP.md, or equivalent per project structure)
    • Any blocking flags, risk register entries, or escalation notes
  2. Classify items into three buckets:

    • Done — completed and verifiably closed (test passing, ticket closed, deployed)
    • Blocked — work that cannot progress without a specific unresolved dependency (name the blocker precisely)
    • Next — the highest-priority items that will be worked next (ordered by priority)
  3. Write the sitrep — structure:

    ## Sitrep — <date> [<scope: sprint / incident / handoff>]
    
    **Done**
    - <item> — <one sentence on what was delivered and its evidence>
    
    **Blocked**
    - <item> — blocked on <specific dependency / owner / ETA if known>
    
    **Next**
    - <item> — <priority order, most important first>
    
    **Risks / flags** (optional — only if material)
    - <risk> — <impact if unresolved> — <owner>
    
  4. Apply the compression test — for each line: if removing it changes nothing that the reader needs to act on, remove it.

  5. Name blockers precisely — "blocked on backend" is not a blocker entry. "Blocked on API contract for /v2/orders — waiting on @owner, ETA unknown" is.

  6. Do not editorialise progress — "made good progress" is noise. "Delivered X, which unblocks Y" is signal.

Outputs

  • A sitrep in the format above
  • Optional: a one-paragraph executive summary for stakeholders who need even less

Guardrails

  • No activity reporting. "Worked on the auth module" is not a sitrep entry. "Auth module: JWT validation complete, token refresh failing under load — investigating" is.
  • Blocked means blocked. If something is moving slowly but not actually blocked, it belongs in "Next" with a note, not "Blocked."
  • Done means done. "95% done" is not done. If it is not verifiably complete, it is "Next."
  • Length is not quality. A five-line sitrep that answers the three questions is better than a twenty-line one that does not.

相关技能