[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ASWRAYP5ANXC98RR3NQFKM":3,"skill-same:01M7ASWRAYP5ANXC98RR3NQFKM":54},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":32,"ownerDescription":6,"updatedAt":33,"install":34,"resources":47},"01M7ASWRAYP5ANXC98RR3NQFKM","small-safe-steps",null,"Small Safe Steps (S3): breaks work into 1-3h increments with zero downtime. Use when asking \"how do I implement\u002Fmigrate\u002Frefactor\", \"what steps to do X\", \"plan safe migration\", or handling risky DB\u002FAPI changes. Applies expand-contract pattern for migrations, refactorings, schema changes.","saski\u002Farnesto","saski","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F140356?v=4","01M7ASBDSNXTRYDYB4K1PEA1RV","01M7ASBDTAKBDAAV99S1GE7HK6",6,"Unlicense","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fsmall-safe-steps\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: small-safe-steps\ndescription: \"Small Safe Steps (S3): breaks work into 1-3h increments with zero downtime. Use when asking \\\"how do I implement\u002Fmigrate\u002Frefactor\\\", \\\"what steps to do X\\\", \\\"plan safe migration\\\", or handling risky DB\u002FAPI changes. Applies expand-contract pattern for migrations, refactorings, schema changes.\"\nallowed-tools:\n  - Read\n  - AskUserQuestion\n---\n\nSTARTER_CHARACTER = 🪜⚡\n\n# Small Safe Steps\n\n## Mission\n\nHelp you divide **ANY work** into the smallest, safest, most valuable steps possible — ideally **1-3 hours each**.\n\n**Core belief:** Risk grows faster than the size of the change.\n\n---\n\n## How to Use This Skill\n\n**Automatic activation - just ask:**\n- \"How do I migrate from SendGrid to AWS SES?\"\n- \"I need to refactor this legacy code\"\n- \"How do I rename this database column safely?\"\n- \"What's the safest way to implement X?\"\n\n**Manual invocation:**\n- `\u002Fsmall-safe-steps` + describe your work\n- Example: `\u002Fsmall-safe-steps Migrate authentication from Auth0 to Keycloak`\n\n**In workflow with other skills:**\n1. Use **story-splitting** → Break large story into smaller ones\n2. Use **complexity-review** → Simplify technical approach\n3. Use **small-safe-steps** (THIS SKILL) → Plan 1-3h implementation steps\n\n---\n\n## When to Use This Skill\n\n**Use when:**\n- Work feels too big or risky (refactoring, migrations, performance fixes)\n- Database schema changes, API changes, service replacements\n- User asks \"how do I implement\" after deciding what to build\n- Work will take more than 1 day\n\n**Do NOT use when:**\n- User is still deciding WHAT to build (use story-splitting or hamburger-method first)\n- Work is already small (\u003C 3 hours)\n- User asks for architectural review (use complexity-review instead)\n\n---\n\n## Applies to ALL Types of Work\n\n### Product Work\n- Feature implementation\n- User stories (after initial splitting)\n- UX improvements\n\n### Technical Work\n- Refactorings\n- Performance improvements\n- Tech debt reduction\n- Architecture changes\n- Database migrations\n\n### Research & Investigation\n- Debugging complex issues\n- Spike\u002FPOC work\n- Technology evaluation\n- Root cause analysis\n\n### Operations & Deployment\n- Infrastructure changes\n- Service migrations\n- Deployment process improvements\n\n---\n\n## Core Process\n\n### Step 1: Understand the Goal\n\nClarify the goal by asking:\n- \"What's the end state we want?\"\n- \"Why are we doing this now?\"\n- \"What's the smallest outcome that would be valuable?\"\n- \"How will we know it's done?\"\n- \"What's the risk if this goes wrong?\"\n\n### Step 2: Detect Risky Changes\n\n🚨 **Identify if this is a risky\u002Fbreaking change:**\n\n**Red flags for risky changes:**\n- Renaming database columns, tables, or fields\n- Changing data types or formats (string → object, XML → JSON)\n- Modifying public APIs or contracts\n- Replacing existing services\u002Flibraries\n- Changes that break backward compatibility\n- Anything that can't be easily reversed\n\n**If risky change detected → Apply Expand-Contract Pattern**\n\nSee \"Quick Reference: Risky Changes\" below and [REFERENCE.md](REFERENCE.md) for detailed examples.\n\n### Step 3: Break Down into Phases\n\nIdentify 3-5 major phases, each independently valuable:\n\n**Example: \"Improve API performance\"**\n- Phase 1: Establish baseline (measure current performance)\n- Phase 2: Identify bottleneck\n- Phase 3: Implement fix\n- Phase 4: Measure improvement\n- Phase 5: (If needed) Fix secondary issues\n\n**Example: \"Migrate from SendGrid to AWS SES\" (risky change)**\n- Phase 1: Expand (add AWS SES alongside SendGrid)\n- Phase 2: Migrate (gradually switch to AWS SES)\n- Phase 3: Contract (remove SendGrid)\n\n### Step 4: Slice Each Phase into 1-3 Hour Steps\n\nFor **EACH phase**, force division into tiny steps.\n\n**Mandatory criteria for each step:**\n- ✅ Takes 1-3 hours maximum\n- ✅ Deployable to production (or creates verifiable artifact)\n- ✅ Reversible (can undo without pain)\n- ✅ Safe (preserves existing functionality)\n- ✅ Testable\n\n**If a step fails these criteria, it's too big.**\n\n### Step 5: Identify Learning vs. Earning Steps\n\nFlag steps by type:\n\n**Learning steps** (time-boxed research\u002Finvestigation):\n- Time-boxed: \"Investigate X (max 2h)\"\n- Goal: Reduce uncertainty, inform decisions\n- Output: Document, decision, or data\n\n**Earning steps** (deliver value):\n- Goal: Ship working software to production\n- Output: Deployed feature, improvement, or fix\n- Measurable user impact\n\n**Key principle:** Learning before earning. Don't build before you understand.\n\n---\n\n## Quick Reference: Risky Changes → Expand-Contract Pattern\n\nUse this table to quickly identify risky changes and apply the correct pattern.\n\n| Change Type | Example | Expand-Contract Phases | Typical Duration |\n|-------------|---------|----------------------|------------------|\n| **Rename DB column** | `email` → `email_address` | 1. Add new column + dual-write\u003Cbr>2. Switch reads to new\u003Cbr>3. Drop old column | 2-3 weeks |\n| **Change data type** | String → JSON object | 1. Add new column + parse\u002Fbackfill\u003Cbr>2. Switch to new format\u003Cbr>3. Drop old column | 2-4 weeks |\n| **API field rename** | `userName` → `username` | 1. Return both fields\u003Cbr>2. Deprecate old, notify consumers\u003Cbr>3. Remove old field | 2-3 months |\n| **Replace service** | SendGrid → AWS SES | 1. Add new service + dual-call\u003Cbr>2. Route traffic % to new\u003Cbr>3. Remove old service | 1 month |\n| **Replace library** | Lodash → Native JS | 1. Add new code alongside old\u003Cbr>2. Migrate callers incrementally\u003Cbr>3. Remove old library | 1-2 weeks |\n| **Refactor logic** | `calculate_v1` → `calculate_v2` | 1. Implement new alongside old\u003Cbr>2. Compare outputs, migrate callers\u003Cbr>3. Remove old function | 1-2 weeks |\n\n**For detailed step-by-step examples, see [REFERENCE.md](REFERENCE.md).**\n\n---\n\n## Expand-Contract Pattern (Summary)\n\nWhen you detect a **risky\u002Fbreaking change**, use the **Expand-Contract pattern** to maintain zero downtime:\n\n### Phase 1: EXPAND (Add New Alongside Old)\n\n**Goal:** System supports BOTH old and new\n\n**Pattern:**\n1. Add new implementation\u002Fcolumn\u002Ffield\u002Fservice alongside the old (1-2h)\n2. Implement dual-write: write to BOTH old and new (2h)\n3. Deploy and verify both paths work in production (1h)\n4. (Optional) Backfill: migrate existing data to new format (1-3h)\n\n**Key principle:** Zero users are affected yet. System continues working with old.\n\n---\n\n### Phase 2: MIGRATE (Switch to New)\n\n**Goal:** Gradually migrate reads\u002Fusage from old to new\n\n**Pattern:**\n1. Update readers\u002Fconsumers to use new path (2h)\n2. Deploy incrementally (feature flag, canary, percentage rollout) (1h)\n3. Monitor for errors, performance issues (passive)\n4. **Keep dual-write active** (safety net to rollback)\n\n**Key principle:** System now uses new, but still maintains old as backup.\n\n---\n\n### Phase 3: CONTRACT (Remove Old)\n\n**Goal:** Clean up by removing old implementation\n\n**Pattern:**\n1. Stop writing to old path (1h)\n2. Deploy and monitor (verify old truly unused) (passive, 1-2 weeks)\n3. Remove old code\u002Fcolumn\u002Fservice (1h)\n4. Clean up migration\u002Fcompatibility code (1h)\n\n**Key principle:** Only remove after old path has ZERO usage for days\u002Fweeks.\n\n---\n\n### Expand-Contract Checklist\n\n**✅ Before MIGRATE phase:**\n- [ ] Dual-write is working (data going to both old and new)\n- [ ] New path is fully functional in production\n- [ ] Monitoring shows both paths are healthy\n- [ ] Rollback plan is clear\n\n**✅ Before CONTRACT phase:**\n- [ ] Old path has ZERO usage (verified via logs\u002Fmonitoring)\n- [ ] New path has been stable for days\u002Fweeks\n- [ ] No errors related to old path\n- [ ] Team agrees it's safe to remove\n\n---\n\n## Techniques by Work Type (Quick Reference)\n\n| Work Type | Primary Pattern | Phases | Key Considerations |\n|-----------|----------------|--------|-------------------|\n| **Refactoring** | Expand-Contract | Add new → Dual-call → Migrate → Remove old | Compare outputs if possible |\n| **Performance** | Baseline → Fix → Verify | Measure → Identify → Implement → Measure again | Always measure before optimizing |\n| **Migration** | Expand-Contract | Add new → Route % → Remove old | Use feature flags for gradual rollout |\n| **Debugging** | Linear investigation | Reproduce → Log → Analyze → Fix → Verify | Time-box investigation steps |\n| **Research\u002FSpike** | Time-boxed learning | Define questions → Evaluate options → Decide | Max 2h per option, document findings |\n| **DB Schema** | Expand-Contract | Add column → Dual-write → Migrate reads → Drop old | Never drop columns immediately |\n| **API Changes** | Expand-Contract | Return both → Deprecate → Remove | Long migration period (months) |\n\n**For detailed examples of each pattern, see [REFERENCE.md](REFERENCE.md).**\n\n---\n\n## Red Flags That a Step is Too Big\n\nWatch for these signs:\n\n- ❌ \"Then we also need to...\"\n- ❌ \"While we're at it, let's...\"\n- ❌ \"This requires...\"\n- ❌ \"First we have to...\"\n- ❌ Multiple verbs in the description (\"implement and test and deploy\")\n- ❌ Takes more than one day\n- ❌ Can't be easily reversed\n- ❌ Affects multiple systems at once\n- ❌ \"We need to coordinate with X team first\"\n\n**When you spot these, force more slicing.**\n\n---\n\n## Coaching Tone\n\n- **Be ruthless about 1-3 hour steps** (no exceptions)\n- **Challenge anything that feels larger**\n- **Detect risky changes proactively** and suggest expand-contract\n- **Ask forcing questions:**\n  - \"What's the smallest thing we could deploy right now?\"\n  - \"Can we learn this before building it?\"\n  - \"What would we do if we had to ship tomorrow?\"\n  - \"How would we roll this back if it fails?\"\n  - \"Is this a breaking change? Should we use expand-contract?\"\n- Use Eduardo Ferro's phrases:\n  - \"What if we only had half the time?\"\n  - \"What's the worst that could happen?\"\n  - \"Can we avoid doing it entirely?\"\n\n---\n\n## Integration with Other Skills\n\nThis skill works in sequence with other skills:\n\n**Typical workflow:**\n1. **story-splitting**: Break down large user stories into smaller ones\n2. **hamburger-method**: Choose vertical slice to implement first\n3. **complexity-review**: Review and simplify technical approach\n4. **micro-steps-coach** (THIS SKILL): Break simplified approach into 1-3h steps\n\n**Use this skill when:**\n- User knows WHAT to build and asks HOW to implement it\n- After architectural decisions are made (complexity-review)\n- When planning execution of a story\u002Ffeature\u002Frefactoring\n\n**Integration examples:**\n- Use **story-splitting** first → \"Admin can create user\" → Then use micro-steps-coach to plan the 1-3h steps\n- Use **hamburger-method** first → Choose slice (manual email notification) → Then use micro-steps-coach for implementation steps\n- Use **complexity-review** first → Simplify to PostgreSQL instead of Kafka → Then use micro-steps-coach for migration steps\n\n**Do NOT use this skill when:**\n- User hasn't decided what to build yet (use story-splitting first)\n- User is proposing complex architecture (use complexity-review first)\n\n---\n\n## Self-Check: Did I Apply This Correctly?\n\nAfter applying this skill, verify:\n\n- [ ] Every step takes 1-3 hours (no step exceeds this)\n- [ ] Every step is deployable to production (or creates verifiable artifact)\n- [ ] Every step is reversible (can undo without major pain)\n- [ ] Risky changes use expand-contract pattern (3 phases: Expand → Migrate → Contract)\n- [ ] I separated learning steps (time-boxed investigation) from earning steps (ship value)\n- [ ] Dual-write is active during ENTIRE migrate phase (not removed too early)\n- [ ] Contract phase only starts after old path has ZERO usage for 1-2 weeks\n- [ ] Each phase is independently valuable (delivers some benefit)\n\n**If any checkbox fails, revisit the breakdown.**\n\n**Red flags that I didn't do this right:**\n- Steps are \"research and implement\" (not separated into learning vs. earning)\n- Steps are \"update database and API\" (too many things at once)\n- Risky change doesn't use expand-contract (trying to rename column in one step)\n- No monitoring\u002Fverification steps between phases\n- Contract phase happens immediately after 100% migration (no waiting period)\n\n---\n\n## Key Principles\n\n1. **Risk grows faster than the size of the change**\n   - Small steps = low risk = fast feedback\n\n2. **Every step must be deployable**\n   - If you can't deploy it, it's too big\n\n3. **Every step must be reversible**\n   - If you can't undo it easily, it's too risky\n\n4. **Zero downtime is non-negotiable**\n   - Use expand-contract for risky changes\n   - System must keep working at all times\n\n5. **Learning before earning**\n   - Investigate before implementing\n   - Time-box research\n   - Don't build without understanding\n\n6. **1-3 hours per step, no exceptions**\n   - If it's bigger, slice more\n   - If you can't slice it, you don't understand it yet\n\n---\n\n## Reference\n\nFor detailed step-by-step examples of expand-contract pattern across different contexts (database changes, API migrations, service replacements, refactorings), see [REFERENCE.md](REFERENCE.md) in this skill directory.\n\nExamples included:\n- Database schema changes (rename column, change data type)\n- API contract changes (rename field, versioning)\n- Service migrations (email provider, authentication, microservices extraction)\n- Refactorings and performance improvements\n- Complete interaction examples with user\n\nAuthor: Eduardo Ferro (expand-contract pattern applied to micro-steps)\nSource: https:\u002F\u002Fwww.eferro.net\u002F\n","STARTER_CHARACTER = 🪜⚡\n\n# Small Safe Steps\n\n## Mission\n\nHelp you divide **ANY work** into the smallest, safest, most valuable steps possible — ideally **1-3 hours each**.\n\n**Core belief:** Risk grows faster than the size of the change.\n\n---\n\n## How to Use This Skill\n\n**Automatic activation - just ask:**\n- \"How do I migrate from SendGrid to AWS SES?\"\n- \"I need to refactor this legacy code\"\n- \"How do I rename this database column safely?\"\n- \"What's the safest way to implement X?\"\n\n**Manual invocation:**\n- `\u002Fsmall-safe-steps` + describe your work\n- Example: `\u002Fsmall-safe-steps Migrate authentication from Auth0 to Keycloak`\n\n**In workflow with other skills:**\n1. Use **story-splitting** → Break large story into smaller ones\n2. Use **complexity-review** → Simplify technical approach\n3. Use **small-safe-steps** (THIS SKILL) → Plan 1-3h implementation steps\n\n---\n\n## When to Use This Skill\n\n**Use when:**\n- Work feels too big or risky (refactoring, migrations, performance fixes)\n- Database schema changes, API changes, service replacements\n- User asks \"how do I implement\" after deciding what to build\n- Work will take more than 1 day\n\n**Do NOT use when:**\n- User is still deciding WHAT to build (use story-splitting or hamburger-method first)\n- Work is already small (\u003C 3 hours)\n- User asks for architectural review (use complexity-review instead)\n\n---\n\n## Applies to ALL Types of Work\n\n### Product Work\n- Feature implementation\n- User stories (after initial splitting)\n- UX improvements\n\n### Technical Work\n- Refactorings\n- Performance improvements\n- Tech debt reduction\n- Architecture changes\n- Database migrations\n\n### Research & Investigation\n- Debugging complex issues\n- Spike\u002FPOC work\n- Technology evaluation\n- Root cause analysis\n\n### Operations & Deployment\n- Infrastructure changes\n- Service migrations\n- Deployment process improvements\n\n---\n\n## Core Process\n\n### Step 1: Understand the Goal\n\nClarify the goal by asking:\n- \"What's the end state we want?\"\n- \"Why are we doing this now?\"\n- \"What's the smallest outcome that would be valuable?\"\n- \"How will we know it's done?\"\n- \"What's the risk if this goes wrong?\"\n\n### Step 2: Detect Risky Changes\n\n🚨 **Identify if this is a risky\u002Fbreaking change:**\n\n**Red flags for risky changes:**\n- Renaming database columns, tables, or fields\n- Changing data types or formats (string → object, XML → JSON)\n- Modifying public APIs or contracts\n- Replacing existing services\u002Flibraries\n- Changes that break backward compatibility\n- Anything that can't be easily reversed\n\n**If risky change detected → Apply Expand-Contract Pattern**\n\nSee \"Quick Reference: Risky Changes\" below and [REFERENCE.md](REFERENCE.md) for detailed examples.\n\n### Step 3: Break Down into Phases\n\nIdentify 3-5 major phases, each independently valuable:\n\n**Example: \"Improve API performance\"**\n- Phase 1: Establish baseline (measure current performance)\n- Phase 2: Identify bottleneck\n- Phase 3: Implement fix\n- Phase 4: Measure improvement\n- Phase 5: (If needed) Fix secondary issues\n\n**Example: \"Migrate from SendGrid to AWS SES\" (risky change)**\n- Phase 1: Expand (add AWS SES alongside SendGrid)\n- Phase 2: Migrate (gradually switch to AWS SES)\n- Phase 3: Contract (remove SendGrid)\n\n### Step 4: Slice Each Phase into 1-3 Hour Steps\n\nFor **EACH phase**, force division into tiny steps.\n\n**Mandatory criteria for each step:**\n- ✅ Takes 1-3 hours maximum\n- ✅ Deployable to production (or creates verifiable artifact)\n- ✅ Reversible (can undo without pain)\n- ✅ Safe (preserves existing functionality)\n- ✅ Testable\n\n**If a step fails these criteria, it's too big.**\n\n### Step 5: Identify Learning vs. Earning Steps\n\nFlag steps by type:\n\n**Learning steps** (time-boxed research\u002Finvestigation):\n- Time-boxed: \"Investigate X (max 2h)\"\n- Goal: Reduce uncertainty, inform decisions\n- Output: Document, decision, or data\n\n**Earning steps** (deliver value):\n- Goal: Ship working software to production\n- Output: Deployed feature, improvement, or fix\n- Measurable user impact\n\n**Key principle:** Learning before earning. Don't build before you understand.\n\n---\n\n## Quick Reference: Risky Changes → Expand-Contract Pattern\n\nUse this table to quickly identify risky changes and apply the correct pattern.\n\n| Change Type | Example | Expand-Contract Phases | Typical Duration |\n|-------------|---------|----------------------|------------------|\n| **Rename DB column** | `email` → `email_address` | 1. Add new column + dual-write\u003Cbr>2. Switch reads to new\u003Cbr>3. Drop old column | 2-3 weeks |\n| **Change data type** | String → JSON object | 1. Add new column + parse\u002Fbackfill\u003Cbr>2. Switch to new format\u003Cbr>3. Drop old column | 2-4 weeks |\n| **API field rename** | `userName` → `username` | 1. Return both fields\u003Cbr>2. Deprecate old, notify consumers\u003Cbr>3. Remove old field | 2-3 months |\n| **Replace service** | SendGrid → AWS SES | 1. Add new service + dual-call\u003Cbr>2. Route traffic % to new\u003Cbr>3. Remove old service | 1 month |\n| **Replace library** | Lodash → Native JS | 1. Add new code alongside old\u003Cbr>2. Migrate callers incrementally\u003Cbr>3. Remove old library | 1-2 weeks |\n| **Refactor logic** | `calculate_v1` → `calculate_v2` | 1. Implement new alongside old\u003Cbr>2. Compare outputs, migrate callers\u003Cbr>3. Remove old function | 1-2 weeks |\n\n**For detailed step-by-step examples, see [REFERENCE.md](REFERENCE.md).**\n\n---\n\n## Expand-Contract Pattern (Summary)\n\nWhen you detect a **risky\u002Fbreaking change**, use the **Expand-Contract pattern** to maintain zero downtime:\n\n### Phase 1: EXPAND (Add New Alongside Old)\n\n**Goal:** System supports BOTH old and new\n\n**Pattern:**\n1. Add new implementation\u002Fcolumn\u002Ffield\u002Fservice alongside the old (1-2h)\n2. Implement dual-write: write to BOTH old and new (2h)\n3. Deploy and verify both paths work in production (1h)\n4. (Optional) Backfill: migrate existing data to new format (1-3h)\n\n**Key principle:** Zero users are affected yet. System continues working with old.\n\n---\n\n### Phase 2: MIGRATE (Switch to New)\n\n**Goal:** Gradually migrate reads\u002Fusage from old to new\n\n**Pattern:**\n1. Update readers\u002Fconsumers to use new path (2h)\n2. Deploy incrementally (feature flag, canary, percentage rollout) (1h)\n3. Monitor for errors, performance issues (passive)\n4. **Keep dual-write active** (safety net to rollback)\n\n**Key principle:** System now uses new, but still maintains old as backup.\n\n---\n\n### Phase 3: CONTRACT (Remove Old)\n\n**Goal:** Clean up by removing old implementation\n\n**Pattern:**\n1. Stop writing to old path (1h)\n2. Deploy and monitor (verify old truly unused) (passive, 1-2 weeks)\n3. Remove old code\u002Fcolumn\u002Fservice (1h)\n4. Clean up migration\u002Fcompatibility code (1h)\n\n**Key principle:** Only remove after old path has ZERO usage for days\u002Fweeks.\n\n---\n\n### Expand-Contract Checklist\n\n**✅ Before MIGRATE phase:**\n- [ ] Dual-write is working (data going to both old and new)\n- [ ] New path is fully functional in production\n- [ ] Monitoring shows both paths are healthy\n- [ ] Rollback plan is clear\n\n**✅ Before CONTRACT phase:**\n- [ ] Old path has ZERO usage (verified via logs\u002Fmonitoring)\n- [ ] New path has been stable for days\u002Fweeks\n- [ ] No errors related to old path\n- [ ] Team agrees it's safe to remove\n\n---\n\n## Techniques by Work Type (Quick Reference)\n\n| Work Type | Primary Pattern | Phases | Key Considerations |\n|-----------|----------------|--------|-------------------|\n| **Refactoring** | Expand-Contract | Add new → Dual-call → Migrate → Remove old | Compare outputs if possible |\n| **Performance** | Baseline → Fix → Verify | Measure → Identify → Implement → Measure again | Always measure before optimizing |\n| **Migration** | Expand-Contract | Add new → Route % → Remove old | Use feature flags for gradual rollout |\n| **Debugging** | Linear investigation | Reproduce → Log → Analyze → Fix → Verify | Time-box investigation steps |\n| **Research\u002FSpike** | Time-boxed learning | Define questions → Evaluate options → Decide | Max 2h per option, document findings |\n| **DB Schema** | Expand-Contract | Add column → Dual-write → Migrate reads → Drop old | Never drop columns immediately |\n| **API Changes** | Expand-Contract | Return both → Deprecate → Remove | Long migration period (months) |\n\n**For detailed examples of each pattern, see [REFERENCE.md](REFERENCE.md).**\n\n---\n\n## Red Flags That a Step is Too Big\n\nWatch for these signs:\n\n- ❌ \"Then we also need to...\"\n- ❌ \"While we're at it, let's...\"\n- ❌ \"This requires...\"\n- ❌ \"First we have to...\"\n- ❌ Multiple verbs in the description (\"implement and test and deploy\")\n- ❌ Takes more than one day\n- ❌ Can't be easily reversed\n- ❌ Affects multiple systems at once\n- ❌ \"We need to coordinate with X team first\"\n\n**When you spot these, force more slicing.**\n\n---\n\n## Coaching Tone\n\n- **Be ruthless about 1-3 hour steps** (no exceptions)\n- **Challenge anything that feels larger**\n- **Detect risky changes proactively** and suggest expand-contract\n- **Ask forcing questions:**\n  - \"What's the smallest thing we could deploy right now?\"\n  - \"Can we learn this before building it?\"\n  - \"What would we do if we had to ship tomorrow?\"\n  - \"How would we roll this back if it fails?\"\n  - \"Is this a breaking change? Should we use expand-contract?\"\n- Use Eduardo Ferro's phrases:\n  - \"What if we only had half the time?\"\n  - \"What's the worst that could happen?\"\n  - \"Can we avoid doing it entirely?\"\n\n---\n\n## Integration with Other Skills\n\nThis skill works in sequence with other skills:\n\n**Typical workflow:**\n1. **story-splitting**: Break down large user stories into smaller ones\n2. **hamburger-method**: Choose vertical slice to implement first\n3. **complexity-review**: Review and simplify technical approach\n4. **micro-steps-coach** (THIS SKILL): Break simplified approach into 1-3h steps\n\n**Use this skill when:**\n- User knows WHAT to build and asks HOW to implement it\n- After architectural decisions are made (complexity-review)\n- When planning execution of a story\u002Ffeature\u002Frefactoring\n\n**Integration examples:**\n- Use **story-splitting** first → \"Admin can create user\" → Then use micro-steps-coach to plan the 1-3h steps\n- Use **hamburger-method** first → Choose slice (manual email notification) → Then use micro-steps-coach for implementation steps\n- Use **complexity-review** first → Simplify to PostgreSQL instead of Kafka → Then use micro-steps-coach for migration steps\n\n**Do NOT use this skill when:**\n- User hasn't decided what to build yet (use story-splitting first)\n- User is proposing complex architecture (use complexity-review first)\n\n---\n\n## Self-Check: Did I Apply This Correctly?\n\nAfter applying this skill, verify:\n\n- [ ] Every step takes 1-3 hours (no step exceeds this)\n- [ ] Every step is deployable to production (or creates verifiable artifact)\n- [ ] Every step is reversible (can undo without major pain)\n- [ ] Risky changes use expand-contract pattern (3 phases: Expand → Migrate → Contract)\n- [ ] I separated learning steps (time-boxed investigation) from earning steps (ship value)\n- [ ] Dual-write is active during ENTIRE migrate phase (not removed too early)\n- [ ] Contract phase only starts after old path has ZERO usage for 1-2 weeks\n- [ ] Each phase is independently valuable (delivers some benefit)\n\n**If any checkbox fails, revisit the breakdown.**\n\n**Red flags that I didn't do this right:**\n- Steps are \"research and implement\" (not separated into learning vs. earning)\n- Steps are \"update database and API\" (too many things at once)\n- Risky change doesn't use expand-contract (trying to rename column in one step)\n- No monitoring\u002Fverification steps between phases\n- Contract phase happens immediately after 100% migration (no waiting period)\n\n---\n\n## Key Principles\n\n1. **Risk grows faster than the size of the change**\n   - Small steps = low risk = fast feedback\n\n2. **Every step must be deployable**\n   - If you can't deploy it, it's too big\n\n3. **Every step must be reversible**\n   - If you can't undo it easily, it's too risky\n\n4. **Zero downtime is non-negotiable**\n   - Use expand-contract for risky changes\n   - System must keep working at all times\n\n5. **Learning before earning**\n   - Investigate before implementing\n   - Time-box research\n   - Don't build without understanding\n\n6. **1-3 hours per step, no exceptions**\n   - If it's bigger, slice more\n   - If you can't slice it, you don't understand it yet\n\n---\n\n## Reference\n\nFor detailed step-by-step examples of expand-contract pattern across different contexts (database changes, API migrations, service replacements, refactorings), see [REFERENCE.md](REFERENCE.md) in this skill directory.\n\nExamples included:\n- Database schema changes (rename column, change data type)\n- API contract changes (rename field, versioning)\n- Service migrations (email provider, authentication, microservices extraction)\n- Refactorings and performance improvements\n- Complete interaction examples with user\n\nAuthor: Eduardo Ferro (expand-contract pattern applied to micro-steps)\nSource: https:\u002F\u002Fwww.eferro.net\u002F\n",{"name":5,"description":7,"allowed-tools":23},[24,25],"Read","AskUserQuestion","main","3c24fc2c035029b08c7326774bee3fef8fb95566",".agents\u002Fskills\u002Fsmall-safe-steps",0,"https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto","https:\u002F\u002Fgithub.com\u002Fsaski","http:\u002F\u002Fsaski.com","2026-11-12 20:09:16",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add saski\u002Farnesto --skill small-safe-steps",{"tool":40,"command":41},"bunx","bunx skills add saski\u002Farnesto --skill small-safe-steps",{"tool":43,"command":44},"pnpm","pnpm dlx skills add saski\u002Farnesto --skill small-safe-steps","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fsmall-safe-steps\u002FSKILL.md",true,[48],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"REFERENCE.md","REFERENCE","参考文档",15341,"https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fsmall-safe-steps\u002FREFERENCE.md",[55,62,69,76,83,90],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":18},"01M7AT7F5FSHNAMFMF01N0QA1X","visualize","Use \u002Fvisualize for creating compact charts, diagrams, or other visuals inline as part of the conversation itself. It chooses among Markdown tables, Mermaid, inline HTML, and other supported in-transcript forms. Use Canvas for durable artifacts outside the transcript that users may revisit, refine, or share.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fvisualize\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":18},"01M7AT79MHSAR3JDDMP7DGERM9","update-cursor-settings","Modify Cursor\u002FVSCode user settings in settings.json. Use when you want to change editor settings, preferences, configuration, themes, font size, tab size, format on save, auto save, keybindings, or any settings.json values.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fupdate-cursor-settings\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":18},"01M7AT73WF3Y22E0FKYAMCCJES","update-cli-config","View and modify Cursor CLI configuration settings in ~\u002F.cursor\u002Fcli-config.json. Use when the user wants to change CLI settings, configure permissions, switch approval mode, enable vim mode, toggle display options, configure sandbox, or manage any CLI preferences.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fupdate-cli-config\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M7AT6F3FBP85AJJ5BJNR8B5W","statusline","Configure a custom status line in the CLI. Use when the user mentions status line, statusline, statusLine, CLI status bar, prompt footer customization, or wants to add session context above the prompt.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fstatusline\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":18},"01M7AT6D24QFS5XRAFN8S5XJ7W","split-to-prs","Split current work into small reviewable PRs. Use when the user asks to split a chat, set of changes, branch, or PR.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fsplit-to-prs\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":18},"01M7AT69FYFN84XZKW452JV5DF","shell","Runs the rest of a \u002Fshell request as a literal shell command. Use only when the user explicitly invokes \u002Fshell and wants the following text executed directly in the terminal.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fshell\u002FSKILL.md",[],[]]