apt-mcp-agent-setup 1.3.10 → 1.3.12

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 (35) hide show
  1. package/bin/cli.js +1 -1
  2. package/bundle/NOTICES.md +0 -0
  3. package/bundle/core.enc +0 -0
  4. package/bundle/mcp-rules.enc +0 -0
  5. package/bundle/skills-aso.enc +0 -0
  6. package/bundle/skills-ba.enc +0 -0
  7. package/bundle/skills-base.enc +0 -0
  8. package/bundle/skills-be.enc +0 -0
  9. package/bundle/skills-design.enc +0 -0
  10. package/bundle/skills-fe.enc +0 -0
  11. package/bundle/skills-mobile.enc +0 -0
  12. package/bundle/skills-pm.enc +0 -0
  13. package/integrity-manifest.json +17 -17
  14. package/package.json +1 -1
  15. package/src/installer/global-setup.js +1 -1
  16. package/src/installer/platform-config.js +1 -1
  17. package/src/installer/prerequisites.js +1 -1
  18. package/src/installer/project-setup.js +1 -1
  19. package/src/installer/setup-wizard.js +1 -1
  20. package/src/license/crypto.js +1 -1
  21. package/src/license/fingerprint.js +1 -1
  22. package/src/license/terms.js +1 -1
  23. package/src/license/verify.js +1 -1
  24. package/src/presets/index.js +1 -1
  25. package/src/proxy/backends.js +1 -1
  26. package/src/proxy/pipeline.js +1 -1
  27. package/src/proxy/router.config.js +1 -1
  28. package/src/proxy/router.js +1 -1
  29. package/src/proxy/server.js +1 -1
  30. package/src/templates/AGENTS.md +21 -0
  31. package/src/templates/CLAUDE.md +50 -20
  32. package/src/templates/GEMINI.md +33 -2
  33. package/src/templates/copilot-instructions.md +32 -3
  34. package/src/templates/cursorrules.mdc +32 -1
  35. package/src/templates/windsurfrules.md +32 -3
package/src/templates/CLAUDE.md CHANGED
@@ -43,27 +43,15 @@ Tool selection rules, index policies, memory conventions, and auto-skill activat
43
43
  - Impact: `impact_analysis` (always update index first)
44
44
  - Memory: `remember` / `recall` / `read_knowledge` (see conventions in mcp-tools.md)
45
45
 
46
- ---
47
-
48
- ## Session Start Protocol (MANDATORY)
46
+ ## ⛔ CRITICAL: apt-mcp-agent TOOL USAGE (P0 — READ FIRST)
49
47
 
50
- At the start of every session:
51
- 1. Read `.agents/ARCHITECTURE.md` to understand Agents, Skills, and Scripts.
52
- 2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
53
- 3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
54
- 4. Call `read_knowledge` to load dynamic knowledge graph from memory.
55
-
56
- ### 🔄 Context Recovery (MANDATORY after any context truncation/compression)
57
- If your conversation context was truncated, compressed, or you are resuming after a checkpoint:
58
- 1. **Call `pipeline_status()` immediately** — it returns recovery instructions and tool reminders
59
- 2. Re-read the session start files above if not already in context
60
- 3. Classify the current user request before taking any code action
48
+ **You have `apt-mcp-agent` MCP tools available. You MUST use them:**
49
+ - **Before ANY code search**: use `explore_code` or `code_search` — NOT grep_search for code symbols
50
+ - **Before ANY code change**: call `pipeline_status` — NO PIPELINE = NO CODE EDITS
51
+ - **After each pipeline step**: call `pipeline_checkpoint` to record progress
52
+ - **Session start**: call `read_knowledge` + `pipeline_status`
61
53
 
62
- ### ⚠️ MCP TOOL ACCESSIBILITY & FALLBACK WARNING (CRITICAL)
63
- If the specialized MCP tool `code_search` or other core MCP tools (e.g., `recall`, `remember`, `impact_analysis`) are not in your available tools list:
64
- - **DO NOT silently fall back to basic text search or skip debugging workflows.**
65
- - **STOP immediately, warn the user** that the `apt-mcp-agent` server did not start successfully or has been disabled in the IDE settings, and recommend they run `npx apt-mcp-agent-setup status` to check their setup.
66
- - Only proceed with fallback tools (e.g. `grep_search`) if the user explicitly approves it.
54
+ Ignoring these tools and using only built-in IDE tools is a **P0 protocol violation**.
67
55
 
68
56
  ### 🛠️ ENFORCED DEVELOPER PIPELINES (MANDATORY — MCP TOOL GATE)
69
57
 
@@ -98,6 +86,28 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
98
86
  - To start a new cycle → call `pipeline_reset()` to clear previous state
99
87
  - The MCP tool enforces ordering at runtime — out-of-order calls are **rejected with error**
100
88
 
89
+ ---
90
+
91
+ ## Session Start Protocol (MANDATORY)
92
+
93
+ At the start of every session:
94
+ 1. Read `.agents/ARCHITECTURE.md` to understand Agents, Skills, and Scripts.
95
+ 2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
96
+ 3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
97
+ 4. Call `read_knowledge` to load dynamic knowledge graph from memory.
98
+ 5. **Read `.agents/skill-triggers.json` to load auto-routing rules for all skills/workflows/tools.**
99
+
100
+ ### 🔄 Context Recovery (MANDATORY after any context truncation/compression)
101
+ If your conversation context was truncated, compressed, or you are resuming after a checkpoint:
102
+ 1. **Call `pipeline_status()` immediately** — it returns recovery instructions and tool reminders
103
+ 2. Re-read the session start files above if not already in context
104
+ 3. Classify the current user request before taking any code action
105
+
106
+ ### ⚠️ MCP TOOL ACCESSIBILITY & FALLBACK WARNING (CRITICAL)
107
+ If the specialized MCP tool `code_search` or other core MCP tools (e.g., `recall`, `remember`, `impact_analysis`) are not in your available tools list:
108
+ - **DO NOT silently fall back to basic text search or skip debugging workflows.**
109
+ - **STOP immediately, warn the user** that the `apt-mcp-agent` server did not start successfully or has been disabled in the IDE settings, and recommend they run `npx apt-mcp-agent-setup status` to check their setup.
110
+ - Only proceed with fallback tools (e.g. `grep_search`) if the user explicitly approves it.
101
111
 
102
112
  ---
103
113
 
@@ -122,7 +132,16 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
122
132
  | `web-mockup-gen` | "mockup web", "design page", web UI concept needed before coding | Load SKILL.md → Generate web design reference images (no code) |
123
133
  | `mobile-mockup-gen` | "mockup app", "design screen", mobile UI concept needed before coding | Load SKILL.md → Generate mobile screen mockups in phone frames |
124
134
  | `mobile-rules` | Any mobile UI implementation: Flutter widget, screen, touch interaction | Load SKILL.md → Apply platform conventions, touch patterns |
125
- | `flutter-*` | Flutter-specific tasks: layout fix, routing, serialization, testing | Auto-select matching flutter skill based on task type |
135
+ | `flutter-fix-layout-issues` | "overflow", "unbounded height", `RenderFlex`, layout constraint error | Load SKILL.md → Fix overflow/constraint issues |
136
+ | `flutter-build-responsive-layout` | "responsive", "tablet", "media query", adaptive layout | Load SKILL.md → Build adaptive layout |
137
+ | `flutter-setup-declarative-routing` | "routing", "deep link", "go_router", URL navigation | Load SKILL.md → Configure declarative routing |
138
+ | `flutter-add-widget-test` | "widget test", "test component", "verify UI rendering" | Load SKILL.md → Write widget test |
139
+ | `flutter-add-integration-test` | "integration test", "e2e", "app test", "Flutter Driver" | Load SKILL.md → Add integration test |
140
+ | `flutter-implement-json-serialization` | "fromJson", "toJson", "model class", "API response mapping" | Load SKILL.md → Create JSON serialization |
141
+ | `flutter-setup-localization` | "localization", "i18n", "intl", "translation", ".arb" | Load SKILL.md → Setup l10n |
142
+ | `flutter-apply-architecture-best-practices` | "architecture", "clean arch", "layer", project structure | Load SKILL.md → Apply layered architecture |
143
+ | `flutter-add-widget-preview` | "preview", "widget catalog", "golden test" | Load SKILL.md → Add widget preview |
144
+ | `flutter-use-http-package` | "http request", "REST API", "fetch data", "GET/POST" | Load SKILL.md → Use http package |
126
145
 
127
146
  ### Productivity & Utility Skills — auto-trigger based on task context:
128
147
 
@@ -143,6 +162,17 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
143
162
  - ❌ Writing UI code without loading design skill = **SLOP RISK**
144
163
  - ❌ Creating mobile screen without `mobile-rules` = **PLATFORM VIOLATION**
145
164
 
165
+ ### ⚠️ DOMAIN SKILL LOADING (CRITICAL — DO NOT SKIP)
166
+
167
+ Before writing ANY UI, design, or visual code:
168
+ 1. **Web UI** → READ `.agents/skills/ui-style-rules/SKILL.md` — MANDATORY
169
+ 2. **Mobile UI** → READ `.agents/skills/mobile-rules/SKILL.md` — MANDATORY
170
+ 3. **Image/asset gen (web)** → READ `.agents/skills/web-mockup-gen/SKILL.md`
171
+ 4. **Image/asset gen (mobile)** → READ `.agents/skills/mobile-mockup-gen/SKILL.md`
172
+ 5. **Architecture decision** → READ `.agents/skills/stress-test-plan/SKILL.md`
173
+
174
+ **Violation = SLOP output. Writing UI code without design skill = P0 violation.**
175
+
146
176
  ---
147
177
 
148
178
  ## Socratic Gate
package/src/templates/GEMINI.md CHANGED
@@ -28,6 +28,16 @@ Single-context layout — one `CONTEXT.md` + `docs/adr/` at the repo root. See `
28
28
 
29
29
  ---
30
30
 
31
+ ## ⛔ CRITICAL: apt-mcp-agent TOOL USAGE (P0 — READ FIRST)
32
+
33
+ **You have `apt-mcp-agent` MCP tools available. You MUST use them:**
34
+ - **Before ANY code search**: use `explore_code` or `code_search` — NOT grep_search for code symbols
35
+ - **Before ANY code change**: call `pipeline_status` — NO PIPELINE = NO CODE EDITS
36
+ - **After each pipeline step**: call `pipeline_checkpoint` to record progress
37
+ - **Session start**: call `read_knowledge` + `pipeline_status`
38
+
39
+ Ignoring these tools and using only built-in IDE tools is a **P0 protocol violation**.
40
+
31
41
  ## CRITICAL: AGENT & SKILL PROTOCOL (START HERE)
32
42
 
33
43
  > **MANDATORY:** You MUST read the appropriate agent file and its skills BEFORE performing any implementation. This is the highest priority rule.
@@ -37,6 +47,7 @@ Single-context layout — one `CONTEXT.md` + `docs/adr/` at the repo root. See `
37
47
  2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
38
48
  3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
39
49
  4. Call `read_knowledge` to load dynamic knowledge graph from memory.
50
+ 5. **Read `.agents/skill-triggers.json` to load auto-routing rules for all skills/workflows/tools.**
40
51
 
41
52
  ### 🔄 Context Recovery (MANDATORY after any context truncation/compression)
42
53
  If your conversation context was truncated, compressed, or you are resuming after a checkpoint:
@@ -141,7 +152,16 @@ UI/UX Design Skills — tự trigger khi touch UI/design:
141
152
  | `web-mockup-gen` | "mockup web", "design page", web UI concept needed before coding | Load SKILL.md → Generate web design reference images (no code) |
142
153
  | `mobile-mockup-gen` | "mockup app", "design screen", mobile UI concept needed before coding | Load SKILL.md → Generate mobile screen mockups in phone frames |
143
154
  | `mobile-rules` | Any mobile UI implementation: Flutter widget, screen, touch interaction | Load SKILL.md → Apply platform conventions, touch patterns, performance |
144
- | `flutter-*` | Flutter-specific tasks: layout fix, routing, serialization, testing, etc. | Auto-select matching flutter skill based on task type |
155
+ | `flutter-fix-layout-issues` | "overflow", "unbounded height", `RenderFlex`, layout error | Load SKILL.md → Fix overflow/constraint issues |
156
+ | `flutter-build-responsive-layout` | "responsive", "tablet", "media query", adaptive layout | Load SKILL.md → Build adaptive layout |
157
+ | `flutter-setup-declarative-routing` | "routing", "deep link", "go_router", URL navigation | Load SKILL.md → Configure declarative routing |
158
+ | `flutter-add-widget-test` | "widget test", "test component", "verify UI rendering" | Load SKILL.md → Write widget test |
159
+ | `flutter-add-integration-test` | "integration test", "e2e", "Flutter Driver" | Load SKILL.md → Add integration test |
160
+ | `flutter-implement-json-serialization` | "fromJson", "toJson", "model class", API mapping | Load SKILL.md → Create JSON serialization |
161
+ | `flutter-setup-localization` | "localization", "i18n", "intl", ".arb" | Load SKILL.md → Setup l10n |
162
+ | `flutter-apply-architecture-best-practices` | "architecture", "layer", project structure | Load SKILL.md → Apply layered architecture |
163
+ | `flutter-add-widget-preview` | "preview", "widget catalog", "golden test" | Load SKILL.md → Add widget preview |
164
+ | `flutter-use-http-package` | "http request", "REST API", "fetch data" | Load SKILL.md → Use http package |
145
165
 
146
166
  Productivity & Utility Skills — tự trigger theo task context:
147
167
 
@@ -159,7 +179,7 @@ Productivity & Utility Skills — tự trigger theo task context:
159
179
  | `draft-shaper` | "shape draft", "edit notes", clean draft to publishable quality | Load SKILL.md → Iterate paragraph by paragraph to edit draft |
160
180
 
161
181
  ### UI/UX Auto-Detection Rules
162
- - **mobile-flutter preset**: Load `mobile-rules` + `ui-style-rules` (adapted) + `mobile-mockup-gen` + matching `flutter-*`
182
+ - **mobile-flutter preset**: Load `mobile-rules` + `ui-style-rules` (adapted) + `mobile-mockup-gen` + matching flutter skill from table above
163
183
  - **fe-react / fe-angular / fe-vue presets**: Load `ui-style-rules` + `web-mockup-gen`
164
184
  - **aso preset**: Load `aso-router` + matching `aso-*` skills
165
185
  - **system preset**: All of the above based on file context
@@ -169,6 +189,17 @@ Productivity & Utility Skills — tự trigger theo task context:
169
189
  - ❌ Creating mobile screen without `mobile-rules` = **PLATFORM VIOLATION**
170
190
  - ❌ Designing web page without `ui-style-rules` = **GENERIC OUTPUT**
171
191
 
192
+ ### ⚠️ DOMAIN SKILL LOADING (CRITICAL — DO NOT SKIP)
193
+
194
+ Before writing ANY UI, design, or visual code:
195
+ 1. **Web UI** → READ `.agents/skills/ui-style-rules/SKILL.md` — MANDATORY
196
+ 2. **Mobile UI** → READ `.agents/skills/mobile-rules/SKILL.md` — MANDATORY
197
+ 3. **Image/asset gen (web)** → READ `.agents/skills/web-mockup-gen/SKILL.md`
198
+ 4. **Image/asset gen (mobile)** → READ `.agents/skills/mobile-mockup-gen/SKILL.md`
199
+ 5. **Architecture decision** → READ `.agents/skills/stress-test-plan/SKILL.md`
200
+
201
+ **Violation = SLOP output. Writing UI code without design skill = P0 violation.**
202
+
172
203
  ---
173
204
 
174
205
  ## 🛑 GLOBAL SOCRATIC GATE (TIER 0)
package/src/templates/copilot-instructions.md CHANGED
@@ -22,6 +22,16 @@ Single-context layout — one `CONTEXT.md` + `docs/adr/` at the repo root. See `
22
22
 
23
23
  ---
24
24
 
25
+ ## ⛔ CRITICAL: apt-mcp-agent TOOL USAGE (P0 — READ FIRST)
26
+
27
+ **You have `apt-mcp-agent` MCP tools available. You MUST use them:**
28
+ - **Before ANY code search**: use `explore_code` or `code_search` — NOT grep_search for code symbols
29
+ - **Before ANY code change**: call `pipeline_status` — NO PIPELINE = NO CODE EDITS
30
+ - **After each pipeline step**: call `pipeline_checkpoint` to record progress
31
+ - **Session start**: call `read_knowledge` + `pipeline_status`
32
+
33
+ Ignoring these tools and using only built-in IDE tools is a **P0 protocol violation**.
34
+
25
35
  ## Session Start Protocol (MANDATORY)
26
36
 
27
37
  At the start of every session:
@@ -29,6 +39,7 @@ At the start of every session:
29
39
  2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
30
40
  3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
31
41
  4. Call `read_knowledge` to load dynamic knowledge graph from memory.
42
+ 5. **Read `.agents/skill-triggers.json` to load auto-routing rules for all skills/workflows/tools.**
32
43
 
33
44
  ### 🔄 Context Recovery (MANDATORY after any context truncation/compression)
34
45
  If your conversation context was truncated, compressed, or you are resuming after a checkpoint:
@@ -117,9 +128,27 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
117
128
  | `web-mockup-gen` | Web UI concept needed before coding | Load SKILL.md → Generate web design reference images |
118
129
  | `mobile-mockup-gen` | Mobile UI concept needed before coding | Load SKILL.md → Generate mobile screen mockups |
119
130
  | `mobile-rules` | Any mobile UI implementation: Flutter, touch interaction | Load SKILL.md → Apply platform conventions |
120
- | `flutter-*` | Flutter-specific tasks: layout, routing, testing | Auto-select matching flutter skill |
121
-
122
- ---
131
+ | `flutter-fix-layout-issues` | "overflow", "unbounded height", `RenderFlex`, layout error | Load SKILL.md → Fix overflow/constraint issues |
132
+ | `flutter-build-responsive-layout` | "responsive", "tablet", "media query", adaptive layout | Load SKILL.md → Build adaptive layout |
133
+ | `flutter-setup-declarative-routing` | "routing", "deep link", "go_router", URL navigation | Load SKILL.md → Configure declarative routing |
134
+ | `flutter-add-widget-test` | "widget test", "test component", "verify UI rendering" | Load SKILL.md → Write widget test |
135
+ | `flutter-add-integration-test` | "integration test", "e2e", "Flutter Driver" | Load SKILL.md → Add integration test |
136
+ | `flutter-implement-json-serialization` | "fromJson", "toJson", "model class", API mapping | Load SKILL.md → Create JSON serialization |
137
+ | `flutter-setup-localization` | "localization", "i18n", "intl", ".arb" | Load SKILL.md → Setup l10n |
138
+ | `flutter-apply-architecture-best-practices` | "architecture", "layer", project structure | Load SKILL.md → Apply layered architecture |
139
+ | `flutter-add-widget-preview` | "preview", "widget catalog", "golden test" | Load SKILL.md → Add widget preview |
140
+ | `flutter-use-http-package` | "http request", "REST API", "fetch data" | Load SKILL.md → Use http package |
141
+
142
+ ### ⚠️ DOMAIN SKILL LOADING (CRITICAL — DO NOT SKIP)
143
+
144
+ Before writing ANY UI, design, or visual code:
145
+ 1. **Web UI** → READ `.agents/skills/ui-style-rules/SKILL.md` — MANDATORY
146
+ 2. **Mobile UI** → READ `.agents/skills/mobile-rules/SKILL.md` — MANDATORY
147
+ 3. **Image/asset gen (web)** → READ `.agents/skills/web-mockup-gen/SKILL.md`
148
+ 4. **Image/asset gen (mobile)** → READ `.agents/skills/mobile-mockup-gen/SKILL.md`
149
+ 5. **Architecture decision** → READ `.agents/skills/stress-test-plan/SKILL.md`
150
+
151
+ **Violation = SLOP output. Writing UI code without design skill = P0 violation.**
123
152
 
124
153
  ## Socratic Gate
125
154
 
package/src/templates/cursorrules.mdc CHANGED
@@ -28,6 +28,16 @@ Single-context layout — one `CONTEXT.md` + `docs/adr/` at the repo root. See `
28
28
 
29
29
  ---
30
30
 
31
+ ## ⛔ CRITICAL: apt-mcp-agent TOOL USAGE (P0 — READ FIRST)
32
+
33
+ **You have `apt-mcp-agent` MCP tools available. You MUST use them:**
34
+ - **Before ANY code search**: use `explore_code` or `code_search` — NOT grep_search for code symbols
35
+ - **Before ANY code change**: call `pipeline_status` — NO PIPELINE = NO CODE EDITS
36
+ - **After each pipeline step**: call `pipeline_checkpoint` to record progress
37
+ - **Session start**: call `read_knowledge` + `pipeline_status`
38
+
39
+ Ignoring these tools and using only built-in IDE tools is a **P0 protocol violation**.
40
+
31
41
  ## 🗺️ Session Start Protocol (MANDATORY)
32
42
 
33
43
  At the start of every session:
@@ -35,6 +45,7 @@ At the start of every session:
35
45
  2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
36
46
  3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
37
47
  4. Call `read_knowledge` to load dynamic knowledge graph from memory.
48
+ 5. **Read `.agents/skill-triggers.json` to load auto-routing rules for all skills/workflows/tools.**
38
49
 
39
50
  ### ⚠️ MCP TOOL ACCESSIBILITY & FALLBACK WARNING (CRITICAL)
40
51
  If the specialized MCP tool `code_search` or other core MCP tools (e.g., `recall`, `remember`, `impact_analysis`) are not in your available tools list:
@@ -116,12 +127,32 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
116
127
  | `web-mockup-gen` | Web UI concept needed before coding | Load SKILL.md → Generate web design reference images |
117
128
  | `mobile-mockup-gen` | Mobile UI concept needed before coding | Load SKILL.md → Generate mobile screen mockups |
118
129
  | `mobile-rules` | Any mobile UI implementation: Flutter, touch interaction | Load SKILL.md → Apply platform conventions |
119
- | `flutter-*` | Flutter-specific tasks: layout, routing, testing | Auto-select matching flutter skill |
130
+ | `flutter-fix-layout-issues` | "overflow", "unbounded height", `RenderFlex`, layout error | Load SKILL.md → Fix overflow/constraint issues |
131
+ | `flutter-build-responsive-layout` | "responsive", "tablet", "media query", adaptive layout | Load SKILL.md → Build adaptive layout |
132
+ | `flutter-setup-declarative-routing` | "routing", "deep link", "go_router", URL navigation | Load SKILL.md → Configure declarative routing |
133
+ | `flutter-add-widget-test` | "widget test", "test component", "verify UI rendering" | Load SKILL.md → Write widget test |
134
+ | `flutter-add-integration-test` | "integration test", "e2e", "Flutter Driver" | Load SKILL.md → Add integration test |
135
+ | `flutter-implement-json-serialization` | "fromJson", "toJson", "model class", API mapping | Load SKILL.md → Create JSON serialization |
136
+ | `flutter-setup-localization` | "localization", "i18n", "intl", ".arb" | Load SKILL.md → Setup l10n |
137
+ | `flutter-apply-architecture-best-practices` | "architecture", "layer", project structure | Load SKILL.md → Apply layered architecture |
138
+ | `flutter-add-widget-preview` | "preview", "widget catalog", "golden test" | Load SKILL.md → Add widget preview |
139
+ | `flutter-use-http-package` | "http request", "REST API", "fetch data" | Load SKILL.md → Use http package |
120
140
 
121
141
  ### Failure Conditions
122
142
  - ❌ Writing UI code without loading design skill = **SLOP RISK**
123
143
  - ❌ Creating mobile screen without `mobile-rules` = **PLATFORM VIOLATION**
124
144
 
145
+ ### ⚠️ DOMAIN SKILL LOADING (CRITICAL — DO NOT SKIP)
146
+
147
+ Before writing ANY UI, design, or visual code:
148
+ 1. **Web UI** → READ `.agents/skills/ui-style-rules/SKILL.md` — MANDATORY
149
+ 2. **Mobile UI** → READ `.agents/skills/mobile-rules/SKILL.md` — MANDATORY
150
+ 3. **Image/asset gen (web)** → READ `.agents/skills/web-mockup-gen/SKILL.md`
151
+ 4. **Image/asset gen (mobile)** → READ `.agents/skills/mobile-mockup-gen/SKILL.md`
152
+ 5. **Architecture decision** → READ `.agents/skills/stress-test-plan/SKILL.md`
153
+
154
+ **Violation = SLOP output. Writing UI code without design skill = P0 violation.**
155
+
125
156
  ---
126
157
 
127
158
  ## 🛑 Socratic Gate
package/src/templates/windsurfrules.md CHANGED
@@ -22,6 +22,16 @@ Single-context layout — one `CONTEXT.md` + `docs/adr/` at the repo root. See `
22
22
 
23
23
  ---
24
24
 
25
+ ## ⛔ CRITICAL: apt-mcp-agent TOOL USAGE (P0 — READ FIRST)
26
+
27
+ **You have `apt-mcp-agent` MCP tools available. You MUST use them:**
28
+ - **Before ANY code search**: use `explore_code` or `code_search` — NOT grep_search for code symbols
29
+ - **Before ANY code change**: call `pipeline_status` — NO PIPELINE = NO CODE EDITS
30
+ - **After each pipeline step**: call `pipeline_checkpoint` to record progress
31
+ - **Session start**: call `read_knowledge` + `pipeline_status`
32
+
33
+ Ignoring these tools and using only built-in IDE tools is a **P0 protocol violation**.
34
+
25
35
  ## Session Start Protocol (MANDATORY)
26
36
 
27
37
  At the start of every session:
@@ -29,6 +39,7 @@ At the start of every session:
29
39
  2. Read `.agents/memory/MEMORY.md` to load persistent project conventions, user preferences, and tech decisions.
30
40
  3. Read `docs/agents/mcp-tools.md` to load MCP tool selection rules & memory conventions.
31
41
  4. Call `read_knowledge` to load dynamic knowledge graph from memory.
42
+ 5. **Read `.agents/skill-triggers.json` to load auto-routing rules for all skills/workflows/tools.**
32
43
 
33
44
  ### 🔄 Context Recovery (MANDATORY after any context truncation/compression)
34
45
  If your conversation context was truncated, compressed, or you are resuming after a checkpoint:
@@ -117,9 +128,27 @@ Skipping steps or executing out-of-order is a **P0 protocol violation** — the
117
128
  | `web-mockup-gen` | Web UI concept needed before coding | Load SKILL.md → Generate web design reference images |
118
129
  | `mobile-mockup-gen` | Mobile UI concept needed before coding | Load SKILL.md → Generate mobile screen mockups |
119
130
  | `mobile-rules` | Any mobile UI implementation: Flutter, touch interaction | Load SKILL.md → Apply platform conventions |
120
- | `flutter-*` | Flutter-specific tasks: layout, routing, testing | Auto-select matching flutter skill |
121
-
122
- ---
131
+ | `flutter-fix-layout-issues` | "overflow", "unbounded height", `RenderFlex`, layout error | Load SKILL.md → Fix overflow/constraint issues |
132
+ | `flutter-build-responsive-layout` | "responsive", "tablet", "media query", adaptive layout | Load SKILL.md → Build adaptive layout |
133
+ | `flutter-setup-declarative-routing` | "routing", "deep link", "go_router", URL navigation | Load SKILL.md → Configure declarative routing |
134
+ | `flutter-add-widget-test` | "widget test", "test component", "verify UI rendering" | Load SKILL.md → Write widget test |
135
+ | `flutter-add-integration-test` | "integration test", "e2e", "Flutter Driver" | Load SKILL.md → Add integration test |
136
+ | `flutter-implement-json-serialization` | "fromJson", "toJson", "model class", API mapping | Load SKILL.md → Create JSON serialization |
137
+ | `flutter-setup-localization` | "localization", "i18n", "intl", ".arb" | Load SKILL.md → Setup l10n |
138
+ | `flutter-apply-architecture-best-practices` | "architecture", "layer", project structure | Load SKILL.md → Apply layered architecture |
139
+ | `flutter-add-widget-preview` | "preview", "widget catalog", "golden test" | Load SKILL.md → Add widget preview |
140
+ | `flutter-use-http-package` | "http request", "REST API", "fetch data" | Load SKILL.md → Use http package |
141
+
142
+ ### ⚠️ DOMAIN SKILL LOADING (CRITICAL — DO NOT SKIP)
143
+
144
+ Before writing ANY UI, design, or visual code:
145
+ 1. **Web UI** → READ `.agents/skills/ui-style-rules/SKILL.md` — MANDATORY
146
+ 2. **Mobile UI** → READ `.agents/skills/mobile-rules/SKILL.md` — MANDATORY
147
+ 3. **Image/asset gen (web)** → READ `.agents/skills/web-mockup-gen/SKILL.md`
148
+ 4. **Image/asset gen (mobile)** → READ `.agents/skills/mobile-mockup-gen/SKILL.md`
149
+ 5. **Architecture decision** → READ `.agents/skills/stress-test-plan/SKILL.md`
150
+
151
+ **Violation = SLOP output. Writing UI code without design skill = P0 violation.**
123
152
 
124
153
  ## Socratic Gate
125
154