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。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/LangeVC/skillweave --skill skillweave-launch命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Create structured PRD (Product Requirements Document) through guided interview. Adapts Ralph Loop concepts for multi-agent AI development. Creates blueprint for promptchain and releasechain workflows.
Multi-Model LLM Council — Deliberation, Peer Review, and Chairman Synthesis with Web Search and Faigate Routing
Analyze design briefs, derive UX principles and design tokens, and evaluate implementation quality.
正在显示 SKILL.md
| facade | true |
| experimental | true |
| name | skillweave-launch |
| description | Coordinate pre-launch checks, environment deployment, communication, verification, and launch metrics. |
| argument-hint | release_summary="[JSON]" command="[deploy|announce|verify|metrics]" environment="[production|staging]" mode="[guided|assisted]" |
Canonical metadata is English. User-facing artifacts follow the output language setting.
Launch coordination — take a published release live.
This skill handles the Launch phase of the SkillWeave lifecycle (phase order: 6). It runs after Release is complete and the artifact is published.
| Activity | Description |
|---|---|
| Production deployment | Coordinate deployment to production environment(s) |
| User communication | Prepare and distribute release notes, changelogs, announcements |
| Go-live coordination | Manage rollout timing, verify deployment health |
| Metrics baseline | Capture pre/post launch metrics for comparison |
Jeder Deployment-Durchlauf MUSS vor Ausführung den Pre-Launch-Check durchlaufen. Bei Fehlschlag wird das Deployment blockiert.
SkillWeave vX.Y.Z — kein zusätzlicher Text im Titel (Beschreibungstext gehört in die Release Notes Body). Regex: ^SkillWeave v[0-9]+\.[0-9]+\.[0-9]+$. Bei Verstoß: Release blockieren.dev → main gemerged (kein direkter Feature-Branch → main Merge). Falls dev nicht existiert: Warnung ausgeben und empfehlen.Jedes Deployment erzeugt einen dokumentierten Rollback-Plan:
{
"git_revert_cmd": "git revert <deploy-hash>",
"db_restore": "backup_<timestamp>.sql",
"estimated_downtime_sec": 30,
"trigger": "health_check.status != 'ok' after deploy"
}
Der Rollback wird NIE automatisch ausgeführt — nur als dokumentierter Plan.
| Aspekt | releasechain | launch |
|---|---|---|
| Phase | Build → Review → Iteration | Deployment → Go-Live |
| Fokus | Code-Qualität, Tests, PRs | Pre-Check, Auslieferung, Metriken |
| Ausführung | Ralph Loop (iterativ) | Linear, gate-basiert |
| Rollback | Code-Revert im Loop | Deployment-Rollback (geplant) |
| Outcome | Fertiges Release-Artifact | Live-System + Announcement |
Invoke the skill by its name with arguments. The skill is host-neutral; no executable prefix is required — route it through any host on any supported transport (Markdown or MCP).
skillweave-launch release_summary='{"version":"0.6.0","artifact_locations":["dist/"],"changelog":"..."}'
release_summary (required): JSON from releasechain containing version, artifact locations, changelogcommand (optional): Select sub-command — deploy, announce, verify, or metricsenvironment (optional): Target environment (default: production)mode (optional): guided (step-by-step) or assisted (automatic with human confirmation at gates)environment (staging/production)deployment.trigger_deployment: Korrekter workflow_dispatch-Aufruf, Environment-Weitergabedeployment.health_check: Positiv (200, <500ms) / Negativ (Timeout, 5xx)deployment.rollback: Plan-Dokumentation, kein automatischer Revertannounce.generate_release_notes: CHANGELOG.md-Parsing, summary-Merge, valides Markdown/JSONannounce.format_announcement: Channel-spezifische Formatierungmetrics.capture_metrics: Endpoint-Abfrage, Timeout-Handlingmetrics.compare_metrics: Delta-Berechnung, Verdict-Logik (Schwellwerte: >5% degraded)