feat(me): add safe automation to launch skill

docs/superpowers/plans/2026-05-13-shipping-launch-safe-automation.md

@@ -0,0 +1,150 @@
+# Shipping Launch Safe Automation Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended)
+> or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax
+> for tracking.
+
+**Goal:** Improve `shipping-and-launch` so agents automatically perform safe launch-preparation work while
+stopping for approval before risky or externally visible actions.
+
+**Architecture:** Modify only the existing `plugins/me/skills/shipping-and-launch/SKILL.md` skill. Add an
+automation policy, execution workflow, decision categories, and `/ship` relationship near the top while
+preserving the existing launch checklist, feature flag, staged rollout, monitoring, and rollback reference
+content.
+
+**Tech Stack:** Markdown skill file with YAML frontmatter, bstack `me` plugin conventions, markdownlint, BATS.
+
+---
+
+## File Structure
+
+- Modify: `plugins/me/skills/shipping-and-launch/SKILL.md` — add safe automation behavior and clarify handoff to `/ship`.
+- No README change planned because the lifecycle summary already describes this skill accurately.
+
+### Task 1: Add safe automation policy
+
+**Files:**
+
+- Modify: `plugins/me/skills/shipping-and-launch/SKILL.md`
+
+- [ ] **Step 1: Read the existing skill**
+
+Run: `sed -n '1,120p' plugins/me/skills/shipping-and-launch/SKILL.md`
+Expected: YAML frontmatter, overview, when-to-use, and pre-launch checklist are visible.
+
+- [ ] **Step 2: Insert automation policy after Overview**
+
+Add a new `## Automation Policy` section after the Overview paragraph. The section must say the agent
+should first do safe, read-only or local reversible work automatically, then stop for approval before risky
+actions.
+
+Include these auto-allowed bullets:
+
+```markdown
+The agent may do these without asking first:
+
+- Inspect git status, diffs, recent commits, changed files, and PR/CI status.
+- Run local read-only or reversible verification such as tests, lint, type checks, builds, audits, and focused smoke checks.
+- Draft launch artifacts: rollout plan, rollback plan, monitoring checklist, post-launch verification checklist, and release notes.
+- Identify missing owners, dashboards, feature flags, environment variables, documentation, or runbooks.
+- Classify findings as `AUTO-COMPLETED`, `NEEDS_APPROVAL`, `BLOCKED`, or `READY_FOR_SHIP_REVIEW`.
+```
+
+Include these approval-required bullets:
+
+```markdown
+The agent must ask before doing these:
+
+- Push branches, create or merge PRs, tag releases, or publish packages.
+- Deploy to staging or production, trigger release workflows, or change infrastructure.
+- Modify feature flags, production configuration, secrets, environment variables, DNS, SSL, or databases.
+- Send external notifications, Slack messages, GitHub comments, status page updates, or customer-facing announcements.
+- Execute rollback, destructive commands, data migrations, or irreversible cleanup.
+```
+
+- [ ] **Step 3: Keep existing checklist content intact**
+
+Do not remove the existing `## The Pre-Launch Checklist` section or its subsections. Only add the new policy before it.
+
+- [ ] **Step 4: Verify policy presence**
+
+Run: `grep -n "Automation Policy\|AUTO-COMPLETED\|NEEDS_APPROVAL\|BLOCKED\|READY_FOR_SHIP_REVIEW\|must ask" plugins/me/skills/shipping-and-launch/SKILL.md`
+Expected: The new section and all decision labels are present.
+
+### Task 2: Add execution workflow and `/ship` handoff
+
+**Files:**
+
+- Modify: `plugins/me/skills/shipping-and-launch/SKILL.md`
+
+- [ ] **Step 1: Insert execution workflow after Automation Policy**
+
+Add a new `## Execution Workflow` section after `## Automation Policy`.
+
+The workflow must contain these steps:
+
+```markdown
+1. Identify the launch type, changed files, blast radius, and whether production systems are affected.
+2. Run safe automatic checks first: local verification, CI/PR status reads, dependency/security audits when available, and documentation checks.
+3. Draft the launch artifacts that can be prepared locally: rollout stages, rollback triggers and procedure, monitoring targets, owners, and post-launch checks.
+4. Categorize every item into `AUTO-COMPLETED`, `NEEDS_APPROVAL`, `BLOCKED`, or `READY_FOR_SHIP_REVIEW`.
+5. Stop on `BLOCKED` items and report the exact evidence.
+6. Ask before any `NEEDS_APPROVAL` action.
+7. When preparation is complete and the change is production-bound, hand off to `/ship` for the final GO/NO-GO decision.
+```
+
+- [ ] **Step 2: Add decision category definitions**
+
+Add a `## Decision Categories` section after the execution workflow.
+
+Use these definitions:
+
+```markdown
+- `AUTO-COMPLETED`: Safe checks or drafts the agent completed locally with evidence.
+- `NEEDS_APPROVAL`: Risky, externally visible, shared-state, or hard-to-reverse actions that require user approval.
+- `BLOCKED`: A launch blocker such as failing tests, missing rollback path, unknown owner, missing monitoring, unresolved security risk, or unverifiable production impact.
+- `READY_FOR_SHIP_REVIEW`: Launch preparation is complete enough for `/ship` to run specialist review and produce GO/NO-GO.
+```
+
+- [ ] **Step 3: Add relationship to `/ship`**
+
+Add a short `## Relationship to /ship` section after decision categories.
+
+The section must say:
+
+```markdown
+Use `shipping-and-launch` to prepare launch artifacts and perform safe automatic checks. Use `/ship` when the question is whether the current change is ready to go live. `/ship` performs the specialist fan-out review and final GO/NO-GO synthesis; this skill prepares the evidence that `/ship` consumes.
+```
+
+- [ ] **Step 4: Verify workflow and handoff presence**
+
+Run: `grep -n "Execution Workflow\|Decision Categories\|Relationship to /ship\|READY_FOR_SHIP_REVIEW\|GO/NO-GO" plugins/me/skills/shipping-and-launch/SKILL.md`
+Expected: The workflow, category definitions, and `/ship` relationship are present.
+
+### Task 3: Validate formatting and repository checks
+
+**Files:**
+
+- Modify: `plugins/me/skills/shipping-and-launch/SKILL.md`
+
+- [ ] **Step 1: Run markdownlint for the changed skill and plan**
+
+Run: `pre-commit run markdownlint --files plugins/me/skills/shipping-and-launch/SKILL.md docs/superpowers/plans/2026-05-13-shipping-launch-safe-automation.md`
+Expected: PASS. If it fails for line length or blank-line formatting, fix only formatting in those files and rerun.
+
+- [ ] **Step 2: Run project tests**
+
+Run: `bats tests/`
+Expected: PASS. If BATS is unavailable or tests fail, report the exact failure and do not claim full verification.
+
+- [ ] **Step 3: Check diff scope**
+
+Run: `git diff -- plugins/me/skills/shipping-and-launch/SKILL.md docs/superpowers/plans/2026-05-13-shipping-launch-safe-automation.md`
+Expected: Diff only contains the shipping-and-launch safe automation policy/workflow and this implementation plan.
+
+## Self-Review
+
+- Spec coverage: The plan covers safe automation, approval gates, execution workflow, decision categories,
+  `/ship` handoff, existing content preservation, and verification.
+- Placeholder scan: No TBD/TODO/later placeholders remain.
+- Type consistency: This plan edits Markdown only; all decision category names are consistent across tasks.

plugins/me/skills/shipping-and-launch/SKILL.md

@@ -7,7 +7,58 @@ description: Prepares production launches. Use when preparing to deploy to produ
 
 ## Overview
 
-Ship with confidence. The goal is not just to deploy — it's to deploy safely, with monitoring in place, a rollback plan ready, and a clear understanding of what success looks like. Every launch should be reversible, observable, and incremental.
+Ship with confidence. The goal is not just to deploy — it's to deploy safely, with monitoring in place,
+a rollback plan ready, and a clear understanding of what success looks like. Every launch should be
+reversible, observable, and incremental.
+
+## Automation Policy
+
+Act first on safe read-only checks and local reversible work. Automatically inspect git status, diffs,
+recent commits, changed files, PR status, and CI status before asking the user for context when those
+checks can answer the question.
+
+Automatically run local verification when appropriate, including tests, lint, type checks, builds,
+audits, and focused smoke checks. Automatically draft rollout plans, rollback plans, monitoring
+checklists, post-launch verification checklists, and release notes for user review.
+
+Automatically identify missing owners, dashboards, feature flags, environment variables, documentation,
+and runbooks. Classify findings as `AUTO-COMPLETED`, `NEEDS_APPROVAL`, `BLOCKED`, or
+`READY_FOR_SHIP_REVIEW` so the user can quickly see what is done, what needs a decision, what is
+blocked, and what is ready for final ship review.
+
+You must ask for approval before any operation that changes shared or external state: push, PR creation,
+merge, release tagging, package publishing, staging or production deploys, release workflows,
+infrastructure changes, feature flag changes, production config changes, secrets, environment variables,
+DNS, SSL, databases, external notifications, Slack messages, GitHub comments, status page updates,
+customer announcements, rollbacks, destructive commands, data migrations, or irreversible cleanup.
+
+## Execution Workflow
+
+1. Identify the launch type, changed files, blast radius, and whether production systems are affected.
+2. Run safe automatic checks first: local verification, CI/PR status reads, dependency/security audits when
+   available, and documentation checks.
+3. Draft launch artifacts that can be prepared locally: rollout stages, rollback triggers/procedure,
+   monitoring targets, owners, and post-launch checks.
+4. Classify every item as `AUTO-COMPLETED`, `NEEDS_APPROVAL`, `BLOCKED`, or `READY_FOR_SHIP_REVIEW`.
+5. If any `BLOCKED` items exist, stop and report the exact evidence.
+6. Ask the user before taking any `NEEDS_APPROVAL` action.
+7. When preparation is complete and the change is production-bound, hand off to `/ship` for the final GO/NO-GO decision.
+
+## Decision Categories
+
+- `AUTO-COMPLETED`: Safe checks or drafts completed locally with evidence.
+- `NEEDS_APPROVAL`: Risky, externally visible, shared-state, or hard-to-reverse actions that require user
+  approval.
+- `BLOCKED`: Launch blocker such as failing tests, missing rollback path, unknown owner, missing
+  monitoring, unresolved security risk, or unverifiable production impact.
+- `READY_FOR_SHIP_REVIEW`: Launch preparation is complete enough for `/ship` to run specialist review and
+  produce GO/NO-GO.
+
+## Relationship to /ship
+
+`shipping-and-launch` prepares launch artifacts and performs safe automatic checks. Use `/ship` when
+deciding whether the current change can go live. `/ship` performs specialist fan-out review and final
+GO/NO-GO synthesis. This skill prepares the evidence that `/ship` consumes.
 
 ## When to Use
 
@@ -93,7 +144,7 @@ return null;
 
 **Feature flag lifecycle:**
 
-```
+```text
 1. DEPLOY with flag OFF     → Code is in production but inactive
 2. ENABLE for team/beta     → Internal testing in production environment
 3. GRADUAL ROLLOUT          → 5% → 25% → 50% → 100% of users
@@ -102,6 +153,7 @@ return null;
 ```
 
 **Rules:**
+
 - Every feature flag has an owner and an expiration date
 - Clean up flags within 2 weeks of full rollout
 - Don't nest feature flags (creates exponential combinations)
@@ -111,7 +163,7 @@ return null;
 
 ### The Rollout Sequence
 
-```
+```text
 1. DEPLOY to staging
    └── Full test suite in staging environment
    └── Manual smoke test of critical flows
@@ -153,6 +205,7 @@ Use these thresholds to decide whether to advance, hold, or roll back at each st
 ### When to Roll Back
 
 Roll back immediately if:
+
 - Error rate increases by more than 2x baseline
 - P95 latency increases by more than 50%
 - User-reported issues spike
@@ -163,7 +216,7 @@ Roll back immediately if:
 
 ### What to Monitor
 
-```
+```text
 Application metrics:
 ├── Error rate (total and by endpoint)
 ├── Response time (p50, p95, p99)
@@ -226,7 +279,7 @@ app.use((err: Error, req: Request, res: Response, next: NextFunction) => {
 
 In the first hour after launch:
 
-```
+```text
 1. Check health endpoint returns 200
 2. Check error monitoring dashboard (no new error types)
 3. Check latency dashboard (no regression)
@@ -263,6 +316,7 @@ Every deployment needs a rollback plan before it happens:
 - Redeploy previous version: < 5 minutes
 - Database rollback: < 15 minutes
 ```
+
 ## See Also
 
 - For security pre-launch checks, see `references/security-checklist.md`