协作头脑风暴
brainstorming
在实现前澄清需求、探索设计选择,并形成清晰的开发方向。
standup【开发与工程】Quick project status report synthesized from state files, artifacts, and git activity. No agent dispatch (<2s).
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Produce a fast, Linear-style project status report by reading the current run state, role artifacts, and git activity. This skill does not dispatch subagents -- the lead agent reads everything directly and synthesizes.
Check for .agenteam/config.yaml, .agenteam.team/config.yaml, or legacy
agenteam.yaml in the project root. If all are missing:
mkdir -p .agenteamcp <plugin-dir>/templates/agenteam.yaml.template .agenteam/config.yamlpython3 <runtime>/agenteam_rt.py generate.agenteam/config.yaml to customize."Call the runtime to assemble run state, health indicator, and artifact paths:
python3 <runtime>/agenteam_rt.py standup
Capture the JSON output. Expected fields:
health -- on-track, at-risk, off-track, or no-active-runrun_id -- current run identifier (may be null if no active run)task -- task descriptionstages -- list of stages with status, assigned role, and gate stateartifact_paths -- map of role name to artifact directorygovernance.adoption -- optional summary of decisions, escalations,
tripwire checks, gate rejections, and criteria overridesoutput_path -- where to write the final report (e.g., docs/meetings/<timestamp>-standup.md)thread_id, last heartbeat and heartbeat age,
idle budget and wall budget remaining, retry count, and stop reasonFor each role in artifact_paths, check what exists in that directory
and produce a one-line summary:
| Role | Artifact Directory | Look For |
|---|---|---|
| researcher | docs/research/ | Research reports, dated findings |
| pm | docs/strategies/ | Strategy docs, roadmaps, specs |
| architect | docs/designs/ | Design docs, architecture decisions |
| dev | src/**, docs/plans/ | Source files changed, plan docs |
| qa | tests/** | Test files, coverage reports |
| reviewer | (read-only) | Review comments in state files |
For each role with artifacts:
If no active run exists, scan artifacts and git history to build a best-effort summary of recent team activity.
Briefly inspect recent git state:
git log --oneline -10
git branch --list
git status --short
Summarize at the changeset level:
Do not include raw commit hashes or full diffs. Report outcomes, not activity.
Produce a Linear-style status report using this format:
Health: [ON TRACK | AT RISK | OFF TRACK]
## Completed
- [role] one-line summary of what's done
## In Progress
- [role] what's happening now
## Blocked
- [role] problem + next step + owner
## Decisions
- key decisions made (from design docs if available)
## Next
- what happens when current work completes
Rules for the report:
health
field. If stages are blocked or gates rejected, it is AT RISK or
OFF TRACK.thread_id, last heartbeat, idle budget, and wall budget for
long-running work.governance.adoption, design docs, or strategy files. Include open
follow-ups, escalations, tripwire blocks, gate rejections, and criteria
overrides when the runtime JSON reports them. Omit this section if none are
found.Write the synthesized report to the output_path from the runtime JSON
(typically docs/meetings/<timestamp>-standup.md):
mkdir -p "$(dirname "$output_path")"
Write the report content to that file.
Show the full report to the user in the conversation. If there is an active run, include the run context header:
AgenTeam Standup: <team-name>
Run: <run-id> | Task: <short task description>
If no active run, use:
AgenTeam Standup: <team-name>
(No active run -- summarizing recent activity)
Resolve the AgenTeam runtime:
./runtime/agenteam_rt.py<plugin-install-path>/runtime/agenteam_rt.pyThis skill should complete in under 2 seconds. It reads local files and runs lightweight git commands -- no LLM subagent calls, no network requests, no web searches.