How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

sync-spec - Prompt - OpenSmartRoute

Name: sync-spec
Author: Minh6625

Prompt file imported from Minh6625/health_system (.windsurf/workflows/sync-spec.md). Copyright stays with the author.

/sync-spec — Spec Drift Detection & Ripple

Spec change without ripple = drift = bugs months later. This workflow systematically traces impact.

Use when:

UC updated (new flow, changed BR, new field).
SRS section revised.
SQL canonical schema changed.
API contract version bumped.

Pre-flight

Invoke skills: karpathy-guidelines (surgical updates), decision-log (if change is architectural).
Identify the spec change: which file, which sections, what semantic change?

Phase 1 — Locate spec change

# What changed in spec?
git -C 'd:\DoAn2\VSmartwatch\PM_REVIEW' diff HEAD~1 -- Resources/UC/<Module>/UC<XXX>.md
git -C 'd:\DoAn2\VSmartwatch\PM_REVIEW' diff HEAD~1 -- "SQL SCRIPTS/"

Classify change type:

Type	Examples	Ripple severity
Behavioral	Main flow step added/changed, BR added	High — code logic + tests
Field added	New data field in UC	High — DB column + model + UI
Field renamed	Field name change	Medium — refactor cascade
Field removed	Deprecated field	High — caller cleanup
NFR change	Performance/security threshold	Variable — review impact
Cosmetic	Wording, formatting	Low — no code impact

Phase 2 — Trace ripple targets

For each spec change, list downstream artifacts to update.

A. Code repos affected

Search for UC reference in code:

foreach ($r in @('HealthGuard','health_system','Iot_Simulator_clean','healthguard-model-api')) {
  Write-Host "=== $r ===" -ForegroundColor Cyan
  Get-ChildItem "d:\DoAn2\VSmartwatch\$r" -Recurse -File -Include '*.dart','*.py','*.js','*.jsx','*.tsx' -ErrorAction SilentlyContinue |
    Select-String -Pattern 'UC<XXX>' -List
}

B. Test files

Tests reference UC behaviors:

# Search for UC in test files
foreach ($r in @('...')) {
  Get-ChildItem "d:\DoAn2\VSmartwatch\$r" -Recurse -File -Include '*test*' |
    Select-String -Pattern 'UC<XXX>' -List
}

C. JIRA Stories

Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\Resources\TASK\JIRA' -Filter 'STORIES.md' -Recurse |
  Select-String -Pattern 'UC<XXX>' -List

D. Related UCs (include/extend chain)

Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\Resources\UC' -Filter '*.md' -Recurse |
  Select-String -Pattern 'UC<XXX>' -List

E. DB schema (if data field change)

Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\SQL SCRIPTS' -Filter '*.sql' |
  Select-String -Pattern '<table_or_column>' -List

F. Test cases in PM_REVIEW

Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\TESTING' -Recurse -Filter '*testcases.md' -ErrorAction SilentlyContinue |
  Select-String -Pattern 'UC<XXX>' -List

Phase 3 — Build ripple plan

Output: actionable todo per affected file.

# Spec Ripple: UC<XXX> change YYYY-MM-DD

## Spec change
- File: `PM_REVIEW/Resources/UC/<Module>/UC<XXX>.md`
- Change: <what changed semantically>
- Type: Behavioral / Field-added / etc.

## Ripple targets

### Code
- [ ] `health_system/lib/features/<area>/<file>.dart:<line>` — update <what>
- [ ] `health_system/backend/app/routers/<file>.py:<line>` — update <what>

### Tests
- [ ] `health_system/test/features/<area>/<file>_test.dart` — add test for new BR
- [ ] `health_system/backend/tests/test_<file>.py` — update assertion

### JIRA Stories
- [ ] `PM_REVIEW/Resources/TASK/JIRA/Sprint-N/<Epic>/STORIES.md` — story `<ID>` references this UC; update acceptance criteria

### Related UCs
- [ ] UC<YYY> includes UC<XXX> — verify still consistent

### DB
- [ ] `PM_REVIEW/SQL SCRIPTS/init_full_setup.sql` — add column
- [ ] `HealthGuard/backend/prisma/schema.prisma` — add field
- [ ] Migration: write `PM_REVIEW/SQL SCRIPTS/migrations/YYYYMMDD_<desc>.sql`

### Test cases
- [ ] `PM_REVIEW/TESTING/<Module>/<Function>_testcases.md` — add new TC for BR

Phase 4 — Decision log (if architectural)

If spec change introduces:

New API contract or major version bump
Breaking change to data model
New cross-repo coordination

→ Write ADR via skill decision-log. Include:

What changed
Why (driving factor)
Migration strategy for existing consumers
Reverse decision triggers

Phase 5 — Cross-repo handoff

If ripple touches ≥ 2 repos → switch to /cross-repo-feature workflow for sequenced execution.

If single repo → /plan to break into vertical-slice tasks, then /build.

Phase 6 — Verify ripple complete

Before claiming "spec synced":

Check	Command
All UC references in code updated	`grep UC<XXX> -r <repo>` shows updated context
Tests reference new behavior	New test exists, runs, passes
JIRA Stories reflect new acceptance	Story acceptance criteria match UC
DB schema matches UC data fields	Manual diff `init_full_setup.sql` vs UC data fields
No orphan references to old field/flow	`grep <old-field>` = empty

Apply skill verification-before-completion.

Phase 7 — Commit

git -C PM_REVIEW commit -m "docs(uc): cập nhật UC<XXX> với <change>

Ripple: <list of repos updated>"

# Per repo
git -C <repo> commit -m "<type>(<scope>): cập nhật theo UC<XXX> change

UC ref: PM_REVIEW@<sha>"

Each commit references the UC commit SHA — traceable across repos.

Anti-patterns

Pattern	Risk
Update UC, skip ripple	Code drifts from spec; bugs surface months later
Update only obvious files	Hidden references in tests/JIRA missed
Skip DB schema sync	Production schema mismatch
Skip ADR for breaking change	Future-you re-debates
Bundle multiple UC changes in one commit	Hard to revert one without others
Skip cross-repo verify	Producer-consumer drift

Output

✅ Ripple plan with all affected files identified
✅ Each affected file actually updated (verified by grep)
✅ Tests updated to reflect new behavior
✅ JIRA Stories updated
✅ DB schema in sync (if data field change)
✅ ADR written if architectural change
✅ Commits reference UC change SHA for traceability

Install sync-spec and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

# /sync-spec — Spec Drift Detection & Ripple > Spec change without ripple = drift = bugs months later. This workflow systematically traces impact. Use when: - UC updated (new flow, changed BR, new field). - SRS section revised. - SQL canonical schema changed. - API contract version bumped. ## Pre-flight 1. **Invoke skills:** `karpathy-guidelines` (surgical updates), `decision-log` (if change is architectural). 2. **Identify the spec change:** which file, which sections, what semantic change? ## Phase 1 — Locate spec change ```pwsh # What changed in spec? git -C 'd:\DoAn2\VSmartwatch\PM_REVIEW' diff HEAD~1 -- Resources/UC/<Module>/UC<XXX>.md git -C 'd:\DoAn2\VSmartwatch\PM_REVIEW' diff HEAD~1 -- "SQL SCRIPTS/" ``` Classify change type: | Type | Examples | Ripple severity | |---|---|---| | **Behavioral** | Main flow step added/changed, BR added | High — code logic + tests | | **Field added** | New data field in UC | High — DB column + model + UI | | **Field renamed** | Field name change | Medium — refactor cascade | | **Field removed** | Deprecated field | High — caller cleanup | | **NFR change** | Performance/security threshold | Variable — review impact | | **Cosmetic** | Wording, formatting | Low — no code impact | ## Phase 2 — Trace ripple targets For each spec change, list downstream artifacts to update. ### A. Code repos affected Search for UC reference in code: ```pwsh foreach ($r in @('HealthGuard','health_system','Iot_Simulator_clean','healthguard-model-api')) { Write-Host "=== $r ===" -ForegroundColor Cyan Get-ChildItem "d:\DoAn2\VSmartwatch\$r" -Recurse -File -Include '*.dart','*.py','*.js','*.jsx','*.tsx' -ErrorAction SilentlyContinue | Select-String -Pattern 'UC<XXX>' -List } ``` ### B. Test files Tests reference UC behaviors: ```pwsh # Search for UC in test files foreach ($r in @('...')) { Get-ChildItem "d:\DoAn2\VSmartwatch\$r" -Recurse -File -Include '*test*' | Select-String -Pattern 'UC<XXX>' -List } ``` ### C. JIRA Stories ```pwsh Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\Resources\TASK\JIRA' -Filter 'STORIES.md' -Recurse | Select-String -Pattern 'UC<XXX>' -List ``` ### D. Related UCs (include/extend chain) ```pwsh Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\Resources\UC' -Filter '*.md' -Recurse | Select-String -Pattern 'UC<XXX>' -List ``` ### E. DB schema (if data field change) ```pwsh Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\SQL SCRIPTS' -Filter '*.sql' | Select-String -Pattern '<table_or_column>' -List ``` ### F. Test cases in PM_REVIEW ```pwsh Get-ChildItem 'd:\DoAn2\VSmartwatch\PM_REVIEW\TESTING' -Recurse -Filter '*testcases.md' -ErrorAction SilentlyContinue | Select-String -Pattern 'UC<XXX>' -List ``` ## Phase 3 — Build ripple plan Output: actionable todo per affected file. ```markdown # Spec Ripple: UC<XXX> change YYYY-MM-DD ## Spec change - File: `PM_REVIEW/Resources/UC/<Module>/UC<XXX>.md` - Change: <what changed semantically> - Type: Behavioral / Field-added / etc. ## Ripple targets ### Code - [ ] `health_system/lib/features/<area>/<file>.dart:<line>` — update <what> - [ ] `health_system/backend/app/routers/<file>.py:<line>` — update <what> ### Tests - [ ] `health_system/test/features/<area>/<file>_test.dart` — add test for new BR - [ ] `health_system/backend/tests/test_<file>.py` — update assertion ### JIRA Stories - [ ] `PM_REVIEW/Resources/TASK/JIRA/Sprint-N/<Epic>/STORIES.md` — story `<ID>` references this UC; update acceptance criteria ### Related UCs - [ ] UC<YYY> includes UC<XXX> — verify still consistent ### DB - [ ] `PM_REVIEW/SQL SCRIPTS/init_full_setup.sql` — add column - [ ] `HealthGuard/backend/prisma/schema.prisma` — add field - [ ] Migration: write `PM_REVIEW/SQL SCRIPTS/migrations/YYYYMMDD_<desc>.sql` ### Test cases - [ ] `PM_REVIEW/TESTING/<Module>/<Function>_testcases.md` — add new TC for BR ``` ## Phase 4 — Decision log (if architectural) If spec change introduces: - New API contract or major version bump - Breaking change to data model - New cross-repo coordination → Write ADR via skill `decision-log`. Include: - What changed - Why (driving factor) - Migration strategy for existing consumers - Reverse decision triggers ## Phase 5 — Cross-repo handoff If ripple touches ≥ 2 repos → switch to `/cross-repo-feature` workflow for sequenced execution. If single repo → `/plan` to break into vertical-slice tasks, then `/build`. ## Phase 6 — Verify ripple complete Before claiming "spec synced": | Check | Command | |---|---| | All UC references in code updated | `grep UC<XXX> -r <repo>` shows updated context | | Tests reference new behavior | New test exists, runs, passes | | JIRA Stories reflect new acceptance | Story acceptance criteria match UC | | DB schema matches UC data fields | Manual diff `init_full_setup.sql` vs UC data fields | | No orphan references to old field/flow | `grep <old-field>` = empty | Apply skill `verification-before-completion`. ## Phase 7 — Commit ```pwsh git -C PM_REVIEW commit -m "docs(uc): cập nhật UC<XXX> với <change> Ripple: <list of repos updated>" # Per repo git -C <repo> commit -m "<type>(<scope>): cập nhật theo UC<XXX> change UC ref: PM_REVIEW@<sha>" ``` Each commit references the UC commit SHA — traceable across repos. ## Anti-patterns | Pattern | Risk | |---|---| | Update UC, skip ripple | Code drifts from spec; bugs surface months later | | Update only obvious files | Hidden references in tests/JIRA missed | | Skip DB schema sync | Production schema mismatch | | Skip ADR for breaking change | Future-you re-debates | | Bundle multiple UC changes in one commit | Hard to revert one without others | | Skip cross-repo verify | Producer-consumer drift | ## Output - ✅ Ripple plan with all affected files identified - ✅ Each affected file actually updated (verified by grep) - ✅ Tests updated to reflect new behavior - ✅ JIRA Stories updated - ✅ DB schema in sync (if data field change) - ✅ ADR written if architectural change - ✅ Commits reference UC change SHA for traceability

{ "prompt": "# /sync-spec — Spec Drift Detection & Ripple\n\n> Spec change without ripple = drift = bugs months later. This workflow systematically traces impact.\n\nUse when:\n- UC updated (new flow, changed BR, new field).\n- SRS section revised.\n- SQL canonical schema changed.\n- API contract version bumped.\n\n## Pre-flight\n\n1. **Invoke skills:** `karpathy-guidelines` (surgical updates), `decision-log` (if change is architectural).\n2. **Identify the spec change:** which file, which sections, what semantic change?\n\n## Phase 1 — Locate spec change\n\n```pwsh\n# What changed in spec?\ngit -C 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW' diff HEAD~1 -- Resources/UC/<Module>/UC<XXX>.md\ngit -C 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW' diff HEAD~1 -- \"SQL SCRIPTS/\"\n```\n\nClassify change type:\n\n| Type | Examples | Ripple severity |\n|---|---|---|\n| **Behavioral** | Main flow step added/changed, BR added | High — code logic + tests |\n| **Field added** | New data field in UC | High — DB column + model + UI |\n| **Field renamed** | Field name change | Medium — refactor cascade |\n| **Field removed** | Deprecated field | High — caller cleanup |\n| **NFR change** | Performance/security threshold | Variable — review impact |\n| **Cosmetic** | Wording, formatting | Low — no code impact |\n\n## Phase 2 — Trace ripple targets\n\nFor each spec change, list downstream artifacts to update.\n\n### A. Code repos affected\n\nSearch for UC reference in code:\n\n```pwsh\nforeach ($r in @('HealthGuard','health_system','Iot_Simulator_clean','healthguard-model-api')) {\n Write-Host \"=== $r ===\" -ForegroundColor Cyan\n Get-ChildItem \"d:\\DoAn2\\VSmartwatch\\$r\" -Recurse -File -Include '*.dart','*.py','*.js','*.jsx','*.tsx' -ErrorAction SilentlyContinue |\n Select-String -Pattern 'UC<XXX>' -List\n}\n```\n\n### B. Test files\n\nTests reference UC behaviors:\n```pwsh\n# Search for UC in test files\nforeach ($r in @('...')) {\n Get-ChildItem \"d:\\DoAn2\\VSmartwatch\\$r\" -Recurse -File -Include '*test*' |\n Select-String -Pattern 'UC<XXX>' -List\n}\n```\n\n### C. JIRA Stories\n\n```pwsh\nGet-ChildItem 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW\\Resources\\TASK\\JIRA' -Filter 'STORIES.md' -Recurse |\n Select-String -Pattern 'UC<XXX>' -List\n```\n\n### D. Related UCs (include/extend chain)\n\n```pwsh\nGet-ChildItem 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW\\Resources\\UC' -Filter '*.md' -Recurse |\n Select-String -Pattern 'UC<XXX>' -List\n```\n\n### E. DB schema (if data field change)\n\n```pwsh\nGet-ChildItem 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW\\SQL SCRIPTS' -Filter '*.sql' |\n Select-String -Pattern '<table_or_column>' -List\n```\n\n### F. Test cases in PM_REVIEW\n\n```pwsh\nGet-ChildItem 'd:\\DoAn2\\VSmartwatch\\PM_REVIEW\\TESTING' -Recurse -Filter '*testcases.md' -ErrorAction SilentlyContinue |\n Select-String -Pattern 'UC<XXX>' -List\n```\n\n## Phase 3 — Build ripple plan\n\nOutput: actionable todo per affected file.\n\n```markdown\n# Spec Ripple: UC<XXX> change YYYY-MM-DD\n\n## Spec change\n- File: `PM_REVIEW/Resources/UC/<Module>/UC<XXX>.md`\n- Change: <what changed semantically>\n- Type: Behavioral / Field-added / etc.\n\n## Ripple targets\n\n### Code\n- [ ] `health_system/lib/features/<area>/<file>.dart:<line>` — update <what>\n- [ ] `health_system/backend/app/routers/<file>.py:<line>` — update <what>\n\n### Tests\n- [ ] `health_system/test/features/<area>/<file>_test.dart` — add test for new BR\n- [ ] `health_system/backend/tests/test_<file>.py` — update assertion\n\n### JIRA Stories\n- [ ] `PM_REVIEW/Resources/TASK/JIRA/Sprint-N/<Epic>/STORIES.md` — story `<ID>` references this UC; update acceptance criteria\n\n### Related UCs\n- [ ] UC<YYY> includes UC<XXX> — verify still consistent\n\n### DB\n- [ ] `PM_REVIEW/SQL SCRIPTS/init_full_setup.sql` — add column\n- [ ] `HealthGuard/backend/prisma/schema.prisma` — add field\n- [ ] Migration: write `PM_REVIEW/SQL SCRIPTS/migrations/YYYYMMDD_<desc>.sql`\n\n### Test cases\n- [ ] `PM_REVIEW/TESTING/<Module>/<Function>_testcases.md` — add new TC for BR\n```\n\n## Phase 4 — Decision log (if architectural)\n\nIf spec change introduces:\n- New API contract or major version bump\n- Breaking change to data model\n- New cross-repo coordination\n\n→ Write ADR via skill `decision-log`. Include:\n- What changed\n- Why (driving factor)\n- Migration strategy for existing consumers\n- Reverse decision triggers\n\n## Phase 5 — Cross-repo handoff\n\nIf ripple touches ≥ 2 repos → switch to `/cross-repo-feature` workflow for sequenced execution.\n\nIf single repo → `/plan` to break into vertical-slice tasks, then `/build`.\n\n## Phase 6 — Verify ripple complete\n\nBefore claiming \"spec synced\":\n\n| Check | Command |\n|---|---|\n| All UC references in code updated | `grep UC<XXX> -r <repo>` shows updated context |\n| Tests reference new behavior | New test exists, runs, passes |\n| JIRA Stories reflect new acceptance | Story acceptance criteria match UC |\n| DB schema matches UC data fields | Manual diff `init_full_setup.sql` vs UC data fields |\n| No orphan references to old field/flow | `grep <old-field>` = empty |\n\nApply skill `verification-before-completion`.\n\n## Phase 7 — Commit\n\n```pwsh\ngit -C PM_REVIEW commit -m \"docs(uc): cập nhật UC<XXX> với <change>\n\nRipple: <list of repos updated>\"\n\n# Per repo\ngit -C <repo> commit -m \"<type>(<scope>): cập nhật theo UC<XXX> change\n\nUC ref: PM_REVIEW@<sha>\"\n```\n\nEach commit references the UC commit SHA — traceable across repos.\n\n## Anti-patterns\n\n| Pattern | Risk |\n|---|---|\n| Update UC, skip ripple | Code drifts from spec; bugs surface months later |\n| Update only obvious files | Hidden references in tests/JIRA missed |\n| Skip DB schema sync | Production schema mismatch |\n| Skip ADR for breaking change | Future-you re-debates |\n| Bundle multiple UC changes in one commit | Hard to revert one without others |\n| Skip cross-repo verify | Producer-consumer drift |\n\n## Output\n\n- ✅ Ripple plan with all affected files identified\n- ✅ Each affected file actually updated (verified by grep)\n- ✅ Tests updated to reflect new behavior\n- ✅ JIRA Stories updated\n- ✅ DB schema in sync (if data field change)\n- ✅ ADR written if architectural change\n- ✅ Commits reference UC change SHA for traceability", "variables": [], "notes": "When UC/SRS/SQL canonical changes, ripple updates across code, tests, JIRA stories, and dependent UCs. Prevents spec drift.", "metadata": { "source": { "provider": "github-windsurf-workflows", "repository": "https://github.com/Minh6625/health_system", "path": ".windsurf/workflows/sync-spec.md", "ref": "14d893863ba1937a069d3c6765d83c73babdc0ff", "url": "https://github.com/Minh6625/health_system/blob/14d893863ba1937a069d3c6765d83c73babdc0ff/.windsurf/workflows/sync-spec.md", "key": "Minh6625/health_system/.windsurf/workflows/sync-spec.md" } } }

sync-spec

About