superkit-mcp-server 1.0.2 → 1.1.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (116) hide show
  1. package/ARCHITECTURE.md +102 -102
  2. package/README.md +67 -63
  3. package/SUPERKIT.md +168 -168
  4. package/agents/code-archaeologist.md +106 -0
  5. package/agents/coder.md +90 -90
  6. package/agents/data-engineer.md +28 -28
  7. package/agents/devops-engineer.md +242 -0
  8. package/agents/git-manager.md +203 -203
  9. package/agents/orchestrator.md +4 -0
  10. package/agents/penetration-tester.md +188 -0
  11. package/agents/performance-optimizer.md +187 -0
  12. package/agents/planner.md +270 -270
  13. package/agents/qa-automation-engineer.md +103 -0
  14. package/agents/quant-developer.md +32 -28
  15. package/agents/reviewer.md +100 -100
  16. package/agents/scout.md +222 -222
  17. package/agents/tester.md +274 -274
  18. package/agents/ui-designer.md +208 -208
  19. package/build/index.js +53 -1
  20. package/build/tools/validators/__tests__/apiSchema.test.js +23 -23
  21. package/build/tools/validators/__tests__/convertRules.test.js +5 -5
  22. package/build/tools/validators/__tests__/frontendDesign.test.js +12 -12
  23. package/build/tools/validators/__tests__/geoChecker.test.js +19 -19
  24. package/build/tools/validators/__tests__/mobileAudit.test.js +12 -12
  25. package/build/tools/validators/__tests__/reactPerformanceChecker.test.js +17 -17
  26. package/build/tools/validators/__tests__/securityScan.test.js +6 -6
  27. package/build/tools/validators/__tests__/seoChecker.test.js +16 -16
  28. package/build/tools/validators/__tests__/typeCoverage.test.js +14 -14
  29. package/package.json +33 -33
  30. package/skills/meta/README.md +30 -30
  31. package/skills/meta/api-design/SKILL.md +134 -134
  32. package/skills/meta/code-review/SKILL.md +44 -37
  33. package/skills/meta/code-review/checklists/pre-merge.md +25 -25
  34. package/skills/meta/code-review/workflows/architecture-pass.md +26 -26
  35. package/skills/meta/code-review/workflows/performance-pass.md +27 -27
  36. package/skills/meta/code-review/workflows/security-pass.md +29 -29
  37. package/skills/meta/compound-docs/SKILL.md +133 -133
  38. package/skills/meta/debug/SKILL.md +40 -40
  39. package/skills/meta/debug/templates/bug-report.template.md +31 -31
  40. package/skills/meta/debug/workflows/reproduce-issue.md +20 -20
  41. package/skills/meta/docker/SKILL.md +126 -126
  42. package/skills/meta/examples/supabase/SKILL.md +46 -46
  43. package/skills/meta/examples/supabase/references/best-practices.md +319 -319
  44. package/skills/meta/examples/supabase/references/common-patterns.md +373 -373
  45. package/skills/meta/examples/supabase/templates/migration-template.sql +49 -49
  46. package/skills/meta/examples/supabase/templates/rls-policy-template.sql +77 -77
  47. package/skills/meta/examples/supabase/workflows/debugging.md +260 -260
  48. package/skills/meta/examples/supabase/workflows/migration-workflow.md +211 -211
  49. package/skills/meta/examples/supabase/workflows/rls-policies.md +244 -244
  50. package/skills/meta/examples/supabase/workflows/schema-design.md +321 -321
  51. package/skills/meta/file-todos/SKILL.md +88 -88
  52. package/skills/meta/mobile/SKILL.md +140 -140
  53. package/skills/meta/nextjs/SKILL.md +101 -101
  54. package/skills/meta/performance/SKILL.md +130 -130
  55. package/skills/meta/react-patterns/SKILL.md +83 -83
  56. package/skills/meta/security/SKILL.md +114 -114
  57. package/skills/meta/session-resume/SKILL.md +96 -96
  58. package/skills/meta/tailwind/SKILL.md +139 -139
  59. package/skills/meta/testing/SKILL.md +43 -43
  60. package/skills/meta/testing/references/vitest-patterns.md +45 -45
  61. package/skills/meta/testing/templates/component-test.template.tsx +37 -37
  62. package/skills/tech/alpha-vantage/SKILL.md +142 -0
  63. package/skills/tech/alpha-vantage/references/commodities.md +153 -0
  64. package/skills/tech/alpha-vantage/references/economic-indicators.md +158 -0
  65. package/skills/tech/alpha-vantage/references/forex-crypto.md +154 -0
  66. package/skills/tech/alpha-vantage/references/fundamentals.md +223 -0
  67. package/skills/tech/alpha-vantage/references/intelligence.md +138 -0
  68. package/skills/tech/alpha-vantage/references/options.md +93 -0
  69. package/skills/tech/alpha-vantage/references/technical-indicators.md +374 -0
  70. package/skills/tech/alpha-vantage/references/time-series.md +157 -0
  71. package/skills/tech/financial-modeling/SKILL.md +18 -0
  72. package/skills/tech/financial-modeling/skills/3-statements/SKILL.md +368 -0
  73. package/skills/tech/financial-modeling/skills/3-statements/references/formatting.md +118 -0
  74. package/skills/tech/financial-modeling/skills/3-statements/references/formulas.md +292 -0
  75. package/skills/tech/financial-modeling/skills/3-statements/references/sec-filings.md +125 -0
  76. package/skills/tech/financial-modeling/skills/dcf-model/SKILL.md +1211 -0
  77. package/skills/tech/financial-modeling/skills/dcf-model/TROUBLESHOOTING.md +40 -0
  78. package/skills/tech/financial-modeling/skills/dcf-model/requirements.txt +8 -0
  79. package/skills/tech/financial-modeling/skills/dcf-model/scripts/validate_dcf.py +292 -0
  80. package/skills/tech/financial-modeling/skills/lbo-model/SKILL.md +236 -0
  81. package/skills/tech/financial-modeling/skills/merger-model/SKILL.md +108 -0
  82. package/skills/tech/intelligent-routing/SKILL.md +5 -5
  83. package/workflows/README.md +191 -191
  84. package/workflows/adr.md +174 -174
  85. package/workflows/changelog.md +74 -74
  86. package/workflows/compound.md +323 -323
  87. package/workflows/compound_health.md +74 -74
  88. package/workflows/create-agent-skill.md +139 -139
  89. package/workflows/cycle.md +144 -144
  90. package/workflows/deploy-docs.md +84 -84
  91. package/workflows/development-rules.md +37 -37
  92. package/workflows/doc.md +95 -95
  93. package/workflows/documentation-management.md +29 -29
  94. package/workflows/explore.md +146 -146
  95. package/workflows/generate_command.md +106 -106
  96. package/workflows/heal-skill.md +97 -97
  97. package/workflows/housekeeping.md +229 -229
  98. package/workflows/kit-setup.md +102 -102
  99. package/workflows/map-codebase.md +78 -0
  100. package/workflows/orchestration-protocol.md +38 -38
  101. package/workflows/plan-compound.md +439 -433
  102. package/workflows/plan_review.md +269 -248
  103. package/workflows/primary-workflow.md +32 -32
  104. package/workflows/promote_pattern.md +86 -86
  105. package/workflows/release-docs.md +82 -82
  106. package/workflows/report-bug.md +135 -135
  107. package/workflows/reproduce-bug.md +118 -118
  108. package/workflows/resolve_pr.md +133 -133
  109. package/workflows/resolve_todo.md +128 -128
  110. package/workflows/review-compound.md +376 -359
  111. package/workflows/skill-review.md +127 -127
  112. package/workflows/specs.md +257 -257
  113. package/workflows/triage-sprint.md +102 -102
  114. package/workflows/triage.md +152 -152
  115. package/workflows/work.md +399 -399
  116. package/workflows/xcode-test.md +93 -93
@@ -1,257 +1,257 @@
1
- ---
2
- description: Create and manage specifications for multi-session initiatives. Use before /plan for major work.
3
- ---
4
-
5
- # /specs - Multi-Session Initiative Management
6
-
7
- Create and manage structured specifications for large, multi-session initiatives like refactors, new features, and rebranding.
8
-
9
- > **Why specs?** Large work spans multiple agent sessions. Specs preserve context, track phases, and capture decisions so each session can resume effectively.
10
-
11
- ## When To Use
12
-
13
- **Use /specs when:**
14
- - Work spans multiple weeks
15
- - Multiple phases with distinct deliverables
16
- - Need to preserve context across agent sessions
17
- - Large refactors, new features, or rebranding
18
-
19
- **Use /plan instead when:**
20
- - Work completes in a single session
21
- - Scope is well-defined and bounded
22
- - No phased rollout needed
23
-
24
- ---
25
-
26
- ## Workflow
27
-
28
- ### Step 0: Search Existing Solutions (MANDATORY)
29
-
30
- > [!CAUTION]
31
- > **BLOCKING STEP.** Before defining a new specification, verify we're not reinventing a solution that already exists in the knowledge base.
32
-
33
- ```bash
34
- // turbo
35
- ./scripts/log-workflow.sh "/specs" "$$"
36
- ./scripts/compound-search.sh "{initiative keywords}"
37
- ```
38
-
39
- **Output required:**
40
- Copy the search output table into your spec's "01-requirements.md" or "02-design.md".
41
-
42
- ---
43
-
44
- ### Step 0.5: Check Existing Specs
45
-
46
- Before creating a new spec, check if one already exists:
47
-
48
- ```bash
49
- ls docs/specs/*/README.md 2>/dev/null
50
- ```
51
-
52
- **If joining an active spec:**
53
- 1. Read `docs/specs/{name}/00-START-HERE.md` for context
54
- 2. Review current phase in `03-tasks.md`
55
- 3. Resume work from where it left off
56
-
57
- **Only proceed to Step 1 if creating a NEW spec.**
58
-
59
- ### Step 1: Create Spec Directory
60
-
61
- ```bash
62
- mkdir -p docs/specs/{name}/{VERIFICATION,plans}
63
- ```
64
-
65
- Use lowercase-hyphenated names: `multi-tenant`, `auth-refactor`, `brand-update`
66
-
67
- ### Step 2: Copy and Customize Templates
68
-
69
- Copy from `docs/specs/templates/`:
70
-
71
- ```bash
72
- cp docs/specs/templates/README-template.md docs/specs/{name}/README.md
73
- cp docs/specs/templates/00-START-HERE-template.md docs/specs/{name}/00-START-HERE.md
74
- cp docs/specs/templates/01-requirements-template.md docs/specs/{name}/01-requirements.md
75
- cp docs/specs/templates/02-design-template.md docs/specs/{name}/02-design.md # Optional for simple specs
76
- cp docs/specs/templates/03-tasks-template.md docs/specs/{name}/03-tasks.md
77
- cp docs/specs/templates/04-decisions-template.md docs/specs/{name}/04-decisions.md
78
- ```
79
-
80
- ### Step 3: Fill Core Documents
81
-
82
- **Priority order:**
83
-
84
- 1. **README.md** - Fill status dashboard with current phase and progress
85
- 2. **01-requirements.md** - Define acceptance criteria for the initiative
86
- 3. **03-tasks.md** - Break work into phases with:
87
- - Clear scope boundaries
88
- - Exit criteria (how to know it's done)
89
- - Estimated duration
90
- - Dependencies on other phases
91
- 4. **00-START-HERE.md** - Summarize current state for future sessions
92
- 5. **04-decisions.md** - Initialize empty (populate as decisions are made)
93
-
94
- ### Step 3.5: Deep Analysis (Think Hard)
95
-
96
- > [!IMPORTANT]
97
- > Specs guide multi-week work. Invest heavily in upfront analysis to avoid costly mid-initiative pivots.
98
-
99
- **Multi-Order Thinking:**
100
- - [ ] **1st order:** What does this initiative directly change?
101
- - [ ] **2nd order:** What systems/processes depend on those changes?
102
- - [ ] **3rd order:** What cascading effects ripple outward?
103
- - [ ] **4th order:** Could this affect unrelated areas through shared dependencies?
104
-
105
- **Long-Term Implications:**
106
- - [ ] How will this age in 6 months? 1 year?
107
- - [ ] Does this reduce or create technical debt?
108
- - [ ] Is this reversible if assumptions prove wrong?
109
-
110
- **Edge Cases (Leave No Stone Unturned):**
111
- - [ ] Boundary conditions and failure modes
112
- - [ ] Concurrent access / race conditions
113
- - [ ] Migration and rollback scenarios
114
-
115
- **Stakeholder Impact Analysis:**
116
- - [ ] **End Users:** Behavior changes? Learning curve?
117
- - [ ] **Other Developers:** Breaking changes? Documentation?
118
- - [ ] **Operations/Support:** New failure modes? Runbooks?
119
- - [ ] **Downstream Systems:** Integration impacts?
120
- - [ ] **Business Stakeholders:** Timeline/resource implications?
121
-
122
- ### Step 4: Create Decision Records
123
-
124
- **For spec-scoped decisions:** Add to `04-decisions.md` (temporary during spec)
125
-
126
- **For project-wide decisions:** Create in `docs/decisions/`
127
- - Technology choices that outlive the spec
128
- - Patterns that become project standards
129
- - Decisions other specs should follow
130
-
131
- **Migration:** When spec completes, promote relevant decisions from `04-decisions.md`
132
- to `docs/decisions/` if they have project-wide applicability.
133
-
134
- ### Step 5: Create Phase 1 Plan
135
-
136
- Run `/plan` for the first phase:
137
-
138
- - Create plan in `docs/specs/{name}/plans/phase1-{description}.md`
139
- - Link back to parent spec in plan header
140
- - Reference phase from `03-tasks.md`
141
-
142
- ### Step 6: Offer Next Steps
143
-
144
- ```
145
- ✓ Spec created: docs/specs/{name}/
146
-
147
- What's next?
148
- 1. Start Phase 1 - Run /plan for first phase
149
- 2. Review spec - Ask for feedback on structure
150
- 3. Share spec - Let team know about the initiative
151
- ```
152
-
153
- ---
154
-
155
- ## Phase Transitions
156
-
157
- When completing a phase:
158
-
159
- 1. **Update 03-tasks.md** - Mark phase tasks complete
160
- 2. **Automate Summary Updates** (MANDATORY):
161
- ```bash
162
- // turbo
163
- ./scripts/update-spec-phase.sh {spec_name} {phase_num} complete
164
- ```
165
- *This atomically updates README.md, START-HERE.md, and 03-tasks.md summary table.*
166
- 3. **Run /plan** for next phase
167
- ```bash
168
- // turbo
169
- ./scripts/update-spec-phase.sh {spec_name} {next_phase_num} started
170
- ```
171
-
172
- ---
173
-
174
- ## Decision Logging
175
-
176
- When making architectural decisions during implementation:
177
-
178
- 1. **Add to 04-decisions.md** using ADR format
179
- 2. **Scope strictly:**
180
- - ✅ Technology choices (GraphQL vs REST)
181
- - ✅ Patterns (Strangler Fig for migration)
182
- - ✅ Trade-offs (speed vs completeness)
183
- - ❌ Bug fixes
184
- - ❌ Implementation details
185
- - ❌ Routine choices
186
-
187
- 3. **Consider pattern promotion:**
188
- ```bash
189
- # If similar decisions made 3+ times
190
- grep -l "{pattern}" docs/specs/*/04-decisions.md | wc -l
191
- ```
192
- If count ≥ 3, run `/compound` to promote to critical pattern.
193
-
194
- ---
195
-
196
- ## Context Restoration Protocol
197
-
198
- Every new session should check for active specs:
199
-
200
- ```bash
201
- # Check for active specs
202
- active_spec=$(ls -d docs/specs/*/ 2>/dev/null | grep -v templates | head -1)
203
- if [ -n "$active_spec" ]; then
204
- echo "Active spec: $active_spec"
205
- cat "${active_spec}00-START-HERE.md"
206
- fi
207
- ```
208
-
209
- This is built into `/plan` Step 0.5 and GEMINI.md agent behavior.
210
-
211
- ---
212
-
213
- ## Quality Guidelines
214
-
215
- **Good specs have:**
216
- - ✅ Clear phase boundaries with exit criteria
217
- - ✅ Acceptance criteria for each requirement
218
- - ✅ Decisions documented with alternatives
219
- - ✅ START-HERE that restores context in 2 minutes
220
-
221
- **Avoid:**
222
- - ❌ Phases without exit criteria
223
- - ❌ Vague requirements without acceptance criteria
224
- - ❌ Decisions without alternatives considered
225
- - ❌ Stale START-HERE that doesn't reflect current state
226
-
227
- ### Phase 5: Completion & Handoff
228
-
229
- #### Step 1: Establish Terminal UI State
230
-
231
- ```javascript
232
- await task_boundary({
233
- TaskName: "[COMPLETED] Spec Management",
234
- TaskStatus: "Spec updated. Offering next steps.",
235
- Mode: "VERIFICATION",
236
- TaskSummary: "Managed spec {name}. Current phase: {phase}. Status: {status}."
237
- });
238
- ```
239
-
240
- #### Step 2: Mandatory Handoff
241
-
242
- ```bash
243
- ✓ Spec managed: docs/specs/{name}/
244
-
245
- Next steps:
246
- 1. /plan - Create plan for next phase
247
- 2. /work - Execute current phase items
248
- ```
249
-
250
- ---
251
-
252
- ## References
253
-
254
- - Templates: `docs/specs/templates/`
255
- - Create phase plans: `/plan`
256
- - Execute phase work: `/work`
257
- - Record solutions: `/compound`
1
+ ---
2
+ description: Create and manage specifications for multi-session initiatives. Use before /plan for major work.
3
+ ---
4
+
5
+ # /specs - Multi-Session Initiative Management
6
+
7
+ Create and manage structured specifications for large, multi-session initiatives like refactors, new features, and rebranding.
8
+
9
+ > **Why specs?** Large work spans multiple agent sessions. Specs preserve context, track phases, and capture decisions so each session can resume effectively.
10
+
11
+ ## When To Use
12
+
13
+ **Use /specs when:**
14
+ - Work spans multiple weeks
15
+ - Multiple phases with distinct deliverables
16
+ - Need to preserve context across agent sessions
17
+ - Large refactors, new features, or rebranding
18
+
19
+ **Use /plan instead when:**
20
+ - Work completes in a single session
21
+ - Scope is well-defined and bounded
22
+ - No phased rollout needed
23
+
24
+ ---
25
+
26
+ ## Workflow
27
+
28
+ ### Step 0: Search Existing Solutions (MANDATORY)
29
+
30
+ > [!CAUTION]
31
+ > **BLOCKING STEP.** Before defining a new specification, verify we're not reinventing a solution that already exists in the knowledge base.
32
+
33
+ ```bash
34
+ // turbo
35
+ ./scripts/log-workflow.sh "/specs" "$$"
36
+ ./scripts/compound-search.sh "{initiative keywords}"
37
+ ```
38
+
39
+ **Output required:**
40
+ Copy the search output table into your spec's "01-requirements.md" or "02-design.md".
41
+
42
+ ---
43
+
44
+ ### Step 0.5: Check Existing Specs
45
+
46
+ Before creating a new spec, check if one already exists:
47
+
48
+ ```bash
49
+ ls docs/specs/*/README.md 2>/dev/null
50
+ ```
51
+
52
+ **If joining an active spec:**
53
+ 1. Read `docs/specs/{name}/00-START-HERE.md` for context
54
+ 2. Review current phase in `03-tasks.md`
55
+ 3. Resume work from where it left off
56
+
57
+ **Only proceed to Step 1 if creating a NEW spec.**
58
+
59
+ ### Step 1: Create Spec Directory
60
+
61
+ ```bash
62
+ mkdir -p docs/specs/{name}/{VERIFICATION,plans}
63
+ ```
64
+
65
+ Use lowercase-hyphenated names: `multi-tenant`, `auth-refactor`, `brand-update`
66
+
67
+ ### Step 2: Copy and Customize Templates
68
+
69
+ Copy from `docs/specs/templates/`:
70
+
71
+ ```bash
72
+ cp docs/specs/templates/README-template.md docs/specs/{name}/README.md
73
+ cp docs/specs/templates/00-START-HERE-template.md docs/specs/{name}/00-START-HERE.md
74
+ cp docs/specs/templates/01-requirements-template.md docs/specs/{name}/01-requirements.md
75
+ cp docs/specs/templates/02-design-template.md docs/specs/{name}/02-design.md # Optional for simple specs
76
+ cp docs/specs/templates/03-tasks-template.md docs/specs/{name}/03-tasks.md
77
+ cp docs/specs/templates/04-decisions-template.md docs/specs/{name}/04-decisions.md
78
+ ```
79
+
80
+ ### Step 3: Fill Core Documents
81
+
82
+ **Priority order:**
83
+
84
+ 1. **README.md** - Fill status dashboard with current phase and progress
85
+ 2. **01-requirements.md** - Define acceptance criteria for the initiative
86
+ 3. **03-tasks.md** - Break work into phases with:
87
+ - Clear scope boundaries
88
+ - Exit criteria (how to know it's done)
89
+ - Estimated duration
90
+ - Dependencies on other phases
91
+ 4. **00-START-HERE.md** - Summarize current state for future sessions
92
+ 5. **04-decisions.md** - Initialize empty (populate as decisions are made)
93
+
94
+ ### Step 3.5: Deep Analysis (Think Hard)
95
+
96
+ > [!IMPORTANT]
97
+ > Specs guide multi-week work. Invest heavily in upfront analysis to avoid costly mid-initiative pivots.
98
+
99
+ **Multi-Order Thinking:**
100
+ - [ ] **1st order:** What does this initiative directly change?
101
+ - [ ] **2nd order:** What systems/processes depend on those changes?
102
+ - [ ] **3rd order:** What cascading effects ripple outward?
103
+ - [ ] **4th order:** Could this affect unrelated areas through shared dependencies?
104
+
105
+ **Long-Term Implications:**
106
+ - [ ] How will this age in 6 months? 1 year?
107
+ - [ ] Does this reduce or create technical debt?
108
+ - [ ] Is this reversible if assumptions prove wrong?
109
+
110
+ **Edge Cases (Leave No Stone Unturned):**
111
+ - [ ] Boundary conditions and failure modes
112
+ - [ ] Concurrent access / race conditions
113
+ - [ ] Migration and rollback scenarios
114
+
115
+ **Stakeholder Impact Analysis:**
116
+ - [ ] **End Users:** Behavior changes? Learning curve?
117
+ - [ ] **Other Developers:** Breaking changes? Documentation?
118
+ - [ ] **Operations/Support:** New failure modes? Runbooks?
119
+ - [ ] **Downstream Systems:** Integration impacts?
120
+ - [ ] **Business Stakeholders:** Timeline/resource implications?
121
+
122
+ ### Step 4: Create Decision Records
123
+
124
+ **For spec-scoped decisions:** Add to `04-decisions.md` (temporary during spec)
125
+
126
+ **For project-wide decisions:** Create in `docs/decisions/`
127
+ - Technology choices that outlive the spec
128
+ - Patterns that become project standards
129
+ - Decisions other specs should follow
130
+
131
+ **Migration:** When spec completes, promote relevant decisions from `04-decisions.md`
132
+ to `docs/decisions/` if they have project-wide applicability.
133
+
134
+ ### Step 5: Create Phase 1 Plan
135
+
136
+ Run `/plan` for the first phase:
137
+
138
+ - Create plan in `docs/specs/{name}/plans/phase1-{description}.md`
139
+ - Link back to parent spec in plan header
140
+ - Reference phase from `03-tasks.md`
141
+
142
+ ### Step 6: Offer Next Steps
143
+
144
+ ```
145
+ ✓ Spec created: docs/specs/{name}/
146
+
147
+ What's next?
148
+ 1. Start Phase 1 - Run /plan for first phase
149
+ 2. Review spec - Ask for feedback on structure
150
+ 3. Share spec - Let team know about the initiative
151
+ ```
152
+
153
+ ---
154
+
155
+ ## Phase Transitions
156
+
157
+ When completing a phase:
158
+
159
+ 1. **Update 03-tasks.md** - Mark phase tasks complete
160
+ 2. **Automate Summary Updates** (MANDATORY):
161
+ ```bash
162
+ // turbo
163
+ ./scripts/update-spec-phase.sh {spec_name} {phase_num} complete
164
+ ```
165
+ *This atomically updates README.md, START-HERE.md, and 03-tasks.md summary table.*
166
+ 3. **Run /plan** for next phase
167
+ ```bash
168
+ // turbo
169
+ ./scripts/update-spec-phase.sh {spec_name} {next_phase_num} started
170
+ ```
171
+
172
+ ---
173
+
174
+ ## Decision Logging
175
+
176
+ When making architectural decisions during implementation:
177
+
178
+ 1. **Add to 04-decisions.md** using ADR format
179
+ 2. **Scope strictly:**
180
+ - ✅ Technology choices (GraphQL vs REST)
181
+ - ✅ Patterns (Strangler Fig for migration)
182
+ - ✅ Trade-offs (speed vs completeness)
183
+ - ❌ Bug fixes
184
+ - ❌ Implementation details
185
+ - ❌ Routine choices
186
+
187
+ 3. **Consider pattern promotion:**
188
+ ```bash
189
+ # If similar decisions made 3+ times
190
+ grep -l "{pattern}" docs/specs/*/04-decisions.md | wc -l
191
+ ```
192
+ If count ≥ 3, run `/compound` to promote to critical pattern.
193
+
194
+ ---
195
+
196
+ ## Context Restoration Protocol
197
+
198
+ Every new session should check for active specs:
199
+
200
+ ```bash
201
+ # Check for active specs
202
+ active_spec=$(ls -d docs/specs/*/ 2>/dev/null | grep -v templates | head -1)
203
+ if [ -n "$active_spec" ]; then
204
+ echo "Active spec: $active_spec"
205
+ cat "${active_spec}00-START-HERE.md"
206
+ fi
207
+ ```
208
+
209
+ This is built into `/plan` Step 0.5 and main AGENT.md (GEMINI.md, CLAUDE.md) agent behavior.
210
+
211
+ ---
212
+
213
+ ## Quality Guidelines
214
+
215
+ **Good specs have:**
216
+ - ✅ Clear phase boundaries with exit criteria
217
+ - ✅ Acceptance criteria for each requirement
218
+ - ✅ Decisions documented with alternatives
219
+ - ✅ START-HERE that restores context in 2 minutes
220
+
221
+ **Avoid:**
222
+ - ❌ Phases without exit criteria
223
+ - ❌ Vague requirements without acceptance criteria
224
+ - ❌ Decisions without alternatives considered
225
+ - ❌ Stale START-HERE that doesn't reflect current state
226
+
227
+ ### Phase 5: Completion & Handoff
228
+
229
+ #### Step 1: Establish Terminal UI State
230
+
231
+ ```javascript
232
+ await task_boundary({
233
+ TaskName: "[COMPLETED] Spec Management",
234
+ TaskStatus: "Spec updated. Offering next steps.",
235
+ Mode: "VERIFICATION",
236
+ TaskSummary: "Managed spec {name}. Current phase: {phase}. Status: {status}."
237
+ });
238
+ ```
239
+
240
+ #### Step 2: Mandatory Handoff
241
+
242
+ ```bash
243
+ ✓ Spec managed: docs/specs/{name}/
244
+
245
+ Next steps:
246
+ 1. /plan - Create plan for next phase
247
+ 2. /work - Execute current phase items
248
+ ```
249
+
250
+ ---
251
+
252
+ ## References
253
+
254
+ - Templates: `docs/specs/templates/`
255
+ - Create phase plans: `/plan`
256
+ - Execute phase work: `/work`
257
+ - Record solutions: `/compound`