菜单
Deployment specialist for the intervention lifecycle. STUB. Packages reviewed artifacts for delivery through the specified channel (SMS vendor API, dashboard endpoint, email system). Parallel to paper's compile+submit. Trigger: deploy, ship, go live, send, /haipipe-application deploy.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
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/jluo41/Tools --skill haipipe-application-deploy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。
基于 SOC 职业分类
正在显示 SKILL.md
| name | haipipe-application-deploy |
| description | Deployment specialist for the intervention lifecycle. STUB. Packages reviewed artifacts for delivery through the specified channel (SMS vendor API, dashboard endpoint, email system). Parallel to paper's compile+submit. Trigger: deploy, ship, go live, send, /haipipe-application deploy. |
| argument-hint | [variant-id] [--channel sms|dashboard|email] [intervention-path] |
| allowed-tools | Bash, Read, Write, Edit, Grep, Glob, Skill |
| metadata | {"version":"0.1.1","last_updated":"2026-07-06","summary":"Deployment specialist (STUB) — packages a reviewed artifact for delivery through its channel. History: ./CHANGELOG.md."} |
Status: STUB. Will package reviewed artifacts for delivery through the specified channel.
When implemented:
Step 1: Verify deployment prerequisites: artifact status reviewed, Gate Ledger
rows complete, venue pack deployment notes satisfied.
Step 2: Package artifact for channel (SMS API payload, dashboard
config, email template).
Step 3: Deploy to staging / test environment.
Step 4: Verify delivery (test send, preview render).
Step 5: On approval: deploy to production.
Step 6: Update artifact status: reviewed → deployed.
Step 7: Log deployment in STATUS.md.
Deployment channels (future):
SMS: vendor API integration (Twilio, etc.)
Dashboard: endpoint update via /haipipe-end
Email: template system integration
In-app: UI component deployment
HIGH — deploys to external channels. Must be gated by review + explicit user approval. Never auto-deploy.
PostFn specialist -- designs/reviews the response-formatting function in an Endpoint_Set. One of 5 inference Fn-types. Called by /haipipe-end when intent references PostFn, response formatting, post-processing, or `post`.
Run any Stage 6 endpoint work: routes intent to the right specialist across Fn-type, artifact-verb, develop-target, and deploy-target axes. Use for designing inference Fns, packaging Endpoint_Sets, training a build, local inference tests, or deploying anywhere. Trigger: endpoint, deploy, develop, train, package, inference Fn, MetaFn, TrigFn, PostFn, Src2InputFn, Input2SrcFn, /haipipe-end.
Src2InputFn specialist -- designs/reviews the record->wire-payload function in an Endpoint_Set (serializes a ProcessedDF record into JSON the model can ingest). Platform-specific: one impl per deploy platform (SageMaker flat JSON vs Databricks dataframe_records); --platform picks (default sagemaker). Called by /haipipe-end when intent references Src2InputFn, record-to-payload serialization, or `src2input`.