@softspark/ai-toolkit 1.3.8 → 1.3.10

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 (26) hide show
  1. package/CHANGELOG.md +22 -0
  2. package/README.md +1 -1
  3. package/app/.claude-plugin/plugin.json +2 -2
  4. package/app/ARCHITECTURE.md +7 -7
  5. package/app/agents/orchestrator.md +21 -3
  6. package/app/rules/claude-toolkit-rules.md +1 -1
  7. package/kb/best-practices/no-hardcoded-counts.md +41 -0
  8. package/kb/reference/agents-catalog.md +3 -3
  9. package/kb/reference/architecture-overview.md +4 -4
  10. package/kb/reference/competitive-features-implementation.md +2 -2
  11. package/kb/reference/manifest-install.md +2 -2
  12. package/kb/reference/skills-catalog.md +3 -3
  13. package/llms-full.txt +64 -44
  14. package/llms.txt +4 -4
  15. package/manifest.json +1 -1
  16. package/package.json +1 -1
  17. package/scripts/generate_augment.py +1 -1
  18. package/scripts/generate_cline.py +1 -1
  19. package/scripts/generate_copilot.py +1 -1
  20. package/scripts/generate_cursor_rules.py +1 -1
  21. package/scripts/generate_gemini.py +4 -4
  22. package/scripts/generate_llms_txt.py +2 -2
  23. package/scripts/generate_windsurf.py +1 -1
  24. package/scripts/generator_base.py +2 -2
  25. package/scripts/install_steps/install_state.py +5 -0
  26. package/kb/reference/commands-catalog.md +0 -21
package/CHANGELOG.md CHANGED
@@ -7,6 +7,28 @@ Versioning follows [Semantic Versioning](https://semver.org/).
7
7
 
8
8
  ---
9
9
 
10
+ ## v1.3.10 — Patch (2026-04-07)
11
+
12
+ ### Fixed
13
+ - **Version cache**: `install` and `update` now clear `~/.ai-toolkit/version-check.json` so `status` shows correct "Latest" immediately after upgrade.
14
+
15
+ ---
16
+
17
+ ## v1.3.9 — Single Source of Truth for Counts (2026-04-07)
18
+
19
+ ### Changed
20
+ - **Counts policy**: Hardcoded skill/agent/hook/test counts removed from all secondary docs. Only README.md, manifest.json, and package.json contain counts. Rule: `kb/best-practices/no-hardcoded-counts.md`.
21
+ - **47 new tests** (327 → 374): MCP manager, language auto-detect, install state, version check, new hooks, orphan cleanup.
22
+
23
+ ### Fixed
24
+ - manifest.json version synced (was 1.0.0)
25
+ - plugin.json version synced (was 1.2.1)
26
+ - agents-catalog.md version aligned
27
+ - ARCHITECTURE.md hook event multipliers corrected
28
+ - Generator scripts no longer embed counts
29
+
30
+ ---
31
+
10
32
  ## v1.3.8 — Language Rules as References (2026-04-07)
11
33
 
12
34
  ### Changed
package/README.md CHANGED
@@ -6,7 +6,7 @@
6
6
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
7
7
  [![Skills](https://img.shields.io/badge/skills-90-brightgreen)](app/skills/)
8
8
  [![Agents](https://img.shields.io/badge/agents-44-blue)](app/agents/)
9
- [![Tests](https://img.shields.io/badge/tests-327%20passing-success)](tests/)
9
+ [![Tests](https://img.shields.io/badge/tests-374%20passing-success)](tests/)
10
10
 
11
11
  ---
12
12
 
package/app/.claude-plugin/plugin.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "ai-toolkit",
3
- "description": "Professional-grade Claude Code toolkit: 90 skills, 44 agents, persona presets, skill security auditor, expanded lifecycle hooks, experimental opt-in plugin packs, benchmark harvesting, and multi-tool support.",
4
- "version": "1.3.8",
3
+ "description": "Professional-grade Claude Code toolkit with persona presets, skill security auditor, expanded lifecycle hooks, experimental opt-in plugin packs, benchmark harvesting, and multi-tool support.",
4
+ "version": "1.3.10",
5
5
  "author": {
6
6
  "name": "SoftSpark",
7
7
  "url": "https://github.com/softspark"
package/app/ARCHITECTURE.md CHANGED
@@ -6,13 +6,13 @@ Universal multi-agent system for software development. Works across all reposito
6
6
 
7
7
  | Component | Count |
8
8
  |-----------|-------|
9
- | Agents | 44 |
10
- | Skills | 90 |
11
- | Hooks | 12 events / 21 entries (SessionStart, Notification, PreToolUse ×2, UserPromptSubmit ×2, PostToolUse, Stop ×2, TaskCompleted, TeammateIdle, SubagentStart, SubagentStop, PreCompact, SessionEnd) |
9
+ | Agents | See agents catalog |
10
+ | Skills | See skills catalog |
11
+ | Hooks | 12 events / 21 entries (SessionStart ×3, Notification ×1, PreToolUse ×4, UserPromptSubmit ×2, PostToolUse ×2, Stop ×2, TaskCompleted ×1, TeammateIdle ×1, SubagentStart ×1, SubagentStop ×1, PreCompact ×2, SessionEnd ×1) |
12
12
 
13
13
  ---
14
14
 
15
- ## Agents (44)
15
+ ## Agents
16
16
 
17
17
  ### Orchestration & Planning (4)
18
18
  | Agent | Model | Purpose |
@@ -100,7 +100,7 @@ Universal multi-agent system for software development. Works across all reposito
100
100
 
101
101
  ---
102
102
 
103
- ## Skills (90)
103
+ ## Skills
104
104
 
105
105
  ### Task Skills (28)
106
106
  | Skill | Slash Command | Purpose |
@@ -255,9 +255,9 @@ Phase 2: IMPLEMENTATION
255
255
  ```
256
256
  .claude/
257
257
  ├── ARCHITECTURE.md # This file
258
- ├── agents/ # Agent definitions (44)
258
+ ├── agents/ # Agent definitions
259
259
  ├── hooks.json # Quality gate hooks (multi-language)
260
- ├── skills/ # All skills: task, hybrid, knowledge (90)
260
+ ├── skills/ # All skills: task, hybrid, knowledge
261
261
  ├── output-styles/ # System prompt output style overrides (e.g. golden-rules)
262
262
  ├── constitution.md # Immutable safety rules (5 articles)
263
263
  └── settings.local.json # Local settings + Agent Teams config
package/app/agents/orchestrator.md CHANGED
@@ -145,7 +145,7 @@ smart_query("[task description in English]") # or hybrid_search_kb()
145
145
  | `data-analyst` | Data | Analysis, visualization, SQL |
146
146
  | `documenter` | Documentation | Architecture notes, runbooks, guides |
147
147
  | `tech-lead` | Architecture | Code review, standards, tech disputes |
148
- | `product-owner` | Product | Requirements, priorities, acceptance criteria |
148
+ | `product-manager` | Product | Requirements, priorities, acceptance criteria |
149
149
  | `security-architect` | Security Design | Threat modeling, secure architecture |
150
150
  | `night-watchman` | Autonomous Ops | Auto-updates, refactoring, cleanup |
151
151
  | `chaos-monkey` | Resilience | Fault injection, reliability testing |
@@ -154,6 +154,24 @@ smart_query("[task description in English]") # or hybrid_search_kb()
154
154
  | `chief-of-staff` | Management | Executive summary, noise filtering |
155
155
  | `meta-architect` | Evolution | Self-optimization, agent updates |
156
156
  | `system-governor` | Safety | Constitution enforcement, VETO power |
157
+ | `ai-engineer` | AI/ML | LLM integration, RAG, vector DBs |
158
+ | `ml-engineer` | ML | Model training, MLOps |
159
+ | `nlp-engineer` | NLP | NER, text classification, NLP pipelines |
160
+ | `llm-ops-engineer` | LLM Ops | LLM caching, fallback, cost optimization |
161
+ | `prompt-engineer` | Prompts | Prompt design, chain-of-thought |
162
+ | `data-scientist` | Data Science | Statistics, EDA, A/B testing |
163
+ | `search-specialist` | Search | Search optimization, relevance tuning |
164
+ | `code-reviewer` | Code Quality | Code review, quality, security |
165
+ | `code-archaeologist` | Legacy Code | Legacy code investigation, dead code |
166
+ | `command-expert` | CLI | Shell scripting, CLI tools |
167
+ | `qa-automation-engineer` | QA | E2E testing, Playwright, Cypress |
168
+ | `mcp-testing-engineer` | MCP Testing | MCP protocol testing |
169
+ | `infrastructure-architect` | Architecture | Architecture decisions, trade-offs |
170
+ | `infrastructure-validator` | Deployment | Deployment verification, health checks |
171
+ | `incident-responder` | Incidents | Production incidents, P1-P4 |
172
+ | `seo-specialist` | SEO | SEO, meta tags, structured data |
173
+ | `fact-checker` | Verification | Claim verification, accuracy |
174
+ | `technical-researcher` | Research | Technical investigation, synthesis |
157
175
 
158
176
  ## 🔴 Agent Boundary Enforcement
159
177
 
@@ -213,7 +231,7 @@ Invoke entire SQUADS in PARALLEL to maximize impact.
213
231
 
214
232
  ### 1. The "Feature Squad" (New Capability)
215
233
  **Target**: Implementing a new user-facing feature.
216
- - `product-owner`: Value & Acceptance Criteria
234
+ - `product-manager`: Value & Acceptance Criteria
217
235
  - `tech-lead`: Architecture & Standards
218
236
  - `frontend-specialist`: UI/UX
219
237
  - `backend-specialist`: Logic & Data
@@ -255,7 +273,7 @@ Invoke entire SQUADS in PARALLEL to maximize impact.
255
273
  □ MCP/Protocol → mcp-specialist
256
274
  □ Discovery → explorer-agent
257
275
  □ Architecture → tech-lead
258
- □ Product → product-owner
276
+ □ Product → product-manager
259
277
  □ Security Design → security-architect
260
278
  □ Maintenance → night-watchman
261
279
  □ Resilience → chaos-monkey
package/app/rules/claude-toolkit-rules.md CHANGED
@@ -1,6 +1,6 @@
1
1
  # Claude Toolkit
2
2
 
3
- Shared AI development toolkit — 90 skills, 44 agents, lifecycle hooks, safety constitution.
3
+ Shared AI development toolkit — lifecycle hooks, safety constitution, multi-platform support.
4
4
 
5
5
  ## Skill Tiers
6
6
 
package/kb/best-practices/no-hardcoded-counts.md ADDED
@@ -0,0 +1,41 @@
1
+ ---
2
+ title: "No Hardcoded Counts in Secondary Docs"
3
+ category: best-practices
4
+ service: ai-toolkit
5
+ tags: [counts, documentation, maintenance, skills, agents, hooks, tests]
6
+ created: "2026-04-07"
7
+ last_updated: "2026-04-07"
8
+ description: "Counts (skills, agents, hooks, tests) should only appear in README.md and manifest.json. All other docs must NOT contain hardcoded numbers to avoid drift."
9
+ ---
10
+
11
+ # No Hardcoded Counts in Secondary Docs
12
+
13
+ ## Rule
14
+
15
+ Hardcoded counts (skills, agents, hooks, tests, plugins) are allowed ONLY in:
16
+ - **README.md** — badges, "What You Get" table, comparison table
17
+ - **manifest.json** — module descriptions
18
+ - **package.json** — description field
19
+
20
+ All other files (CLAUDE.md, ARCHITECTURE.md, KB docs, plugin.json, llms.txt, AGENTS.md, copilot-instructions, rules, GEMINI.md) must NOT contain hardcoded counts like "90 skills" or "44 agents".
21
+
22
+ ## Why
23
+
24
+ Every time a skill, agent, or hook is added/removed, dozens of files need updating. This causes constant drift and stale counts that erode trust. Consolidating to 2-3 files makes maintenance tractable.
25
+
26
+ ## How to Apply
27
+
28
+ - In secondary docs, use relative language: "all agents", "the full skill set", "available hooks"
29
+ - If a doc MUST reference scale, use: "see README.md for current counts"
30
+ - `validate.py` checks counts only in README.md badges — that's sufficient
31
+ - When adding skills/agents/hooks, update ONLY: README.md badges + manifest.json descriptions
32
+
33
+ ## Anti-Pattern
34
+
35
+ ```markdown
36
+ # BAD — hardcoded count in ARCHITECTURE.md
37
+ Shared AI development toolkit — 90 skills, 44 agents
38
+
39
+ # GOOD — no count
40
+ Shared AI development toolkit with multi-platform support
41
+ ```
package/kb/reference/agents-catalog.md CHANGED
@@ -3,13 +3,13 @@ title: "AI Toolkit - Agents Catalog"
3
3
  category: reference
4
4
  service: ai-toolkit
5
5
  tags: [agents, catalog, roles, ai-development]
6
- version: "2.0.0"
6
+ version: "1.3.10"
7
7
  created: "2026-03-23"
8
8
  last_updated: "2026-03-25"
9
- description: "Complete catalog of 44 specialized agents with roles, models, and use cases."
9
+ description: "Complete catalog of specialized agents with roles, models, and use cases."
10
10
  ---
11
11
 
12
- # Agents Catalog (44 agents)
12
+ # Agents Catalog
13
13
 
14
14
  ## By Category
15
15
 
package/kb/reference/architecture-overview.md CHANGED
@@ -3,7 +3,7 @@ title: "AI Toolkit - Architecture Overview"
3
3
  category: reference
4
4
  service: ai-toolkit
5
5
  tags: [architecture, overview, design, structure]
6
- version: "1.0.0"
6
+ version: "1.3.10"
7
7
  created: "2026-03-23"
8
8
  last_updated: "2026-04-07"
9
9
  description: "Architecture of ai-toolkit: directory layout, global install model, skill tiers, and integration with projects."
@@ -13,7 +13,7 @@ description: "Architecture of ai-toolkit: directory layout, global install model
13
13
 
14
14
  ## Purpose
15
15
 
16
- Shared, project-agnostic AI development toolkit for Claude Code (and compatible assistants like Cursor, Windsurf, Copilot, Gemini, Cline, Roo Code, Aider, and Augment). Provides 44 specialized agents, 90 skills (slash commands + knowledge), expanded lifecycle hooks, persona presets, and experimental opt-in plugin packs that teams can adopt separately from the default global install.
16
+ Shared, project-agnostic AI development toolkit for Claude Code (and compatible assistants like Cursor, Windsurf, Copilot, Gemini, Cline, Roo Code, Aider, and Augment). Provides specialized agents, skills (slash commands + knowledge), expanded lifecycle hooks, persona presets, and experimental opt-in plugin packs that teams can adopt separately from the default global install.
17
17
 
18
18
  ## Design Principles
19
19
 
@@ -30,8 +30,8 @@ ai-toolkit/
30
30
  bin/
31
31
  ai-toolkit.js # CLI entry point (install, init, add-rule, ...)
32
32
  app/ # All toolkit components
33
- agents/ # 44 agent definitions (.md + YAML frontmatter)
34
- skills/ # 90 skills: task, hybrid, knowledge
33
+ agents/ # Agent definitions (.md + YAML frontmatter)
34
+ skills/ # skills: task, hybrid, knowledge
35
35
  rules/ # Rules auto-injected into ~/.claude/CLAUDE.md
36
36
  hooks/ # Hook scripts (copied to ~/.ai-toolkit/hooks/)
37
37
  hooks.json # Hook definitions (merged into ~/.claude/settings.json)
package/kb/reference/competitive-features-implementation.md CHANGED
@@ -35,7 +35,7 @@ Strengthen ai-toolkit's competitive position by implementing 10 features from co
35
35
  **Key design principle:** ai-toolkit is a **generic toolkit** — it does NOT know about rag-mcp or any specific consumer. Consumers (like rag-mcp) use ai-toolkit's public API (`inject-rule`, `inject-hook`, `merge-hooks`) to add their own rules and hooks.
36
36
 
37
37
  **State before plan:** 88 skills, 47 agents, 14 hooks, 9 editor integrations
38
- **State after plan:** 90 skills, 44 agents, 21 hooks, 70 language rules, 25 MCP templates, extension API
38
+ **State after plan:** See manifest.json for current counts. Target: skills, agents, hooks, 70 language rules, 25 MCP templates, extension API
39
39
 
40
40
  ---
41
41
 
@@ -384,7 +384,7 @@ All entries are tagged with "_source": "<source-name>" for idempotent updates.
384
384
  "required": true
385
385
  },
386
386
  "agents": {
387
- "description": "44 specialized agents",
387
+ "description": "Specialized agents",
388
388
  "files": ["app/agents/*.md"],
389
389
  "required": false,
390
390
  "default": true
package/kb/reference/manifest-install.md CHANGED
@@ -24,8 +24,8 @@ Modules are defined in `manifest.json` at the repository root. There are 17 modu
24
24
  | Module | Description | In Profile |
25
25
  |--------|-------------|-----------|
26
26
  | `core` | Core hooks and essential skills | minimal, standard, strict, full |
27
- | `agents` | 44 specialized agents | standard, strict, full |
28
- | `skills` | 90 skills (task, hybrid, knowledge) | standard, strict, full |
27
+ | `agents` | Specialized agents | standard, strict, full |
28
+ | `skills` | All skills (task, hybrid, knowledge) | standard, strict, full |
29
29
  | `rules-common` | Common coding rules (5 files) | standard, strict, full |
30
30
  | `rules-typescript` | TypeScript-specific rules (5 files) | auto-detect |
31
31
  | `rules-python` | Python-specific rules (5 files) | auto-detect |
package/kb/reference/skills-catalog.md CHANGED
@@ -3,13 +3,13 @@ title: "AI Toolkit - Skills Catalog"
3
3
  category: reference
4
4
  service: ai-toolkit
5
5
  tags: [skills, domain-knowledge, catalog, task-skills, hybrid-skills]
6
- version: "1.3.8"
6
+ version: "1.3.10"
7
7
  created: "2026-03-23"
8
8
  last_updated: "2026-04-07"
9
- description: "Complete catalog of 90 skills: 28 task, 30 hybrid, 32 knowledge. Includes effort levels, skill-scoped hooks, executable scripts, security auditor, and persona presets."
9
+ description: "Complete skills catalog with task, hybrid, and knowledge skills. Includes effort levels, skill-scoped hooks, executable scripts, security auditor, and persona presets."
10
10
  ---
11
11
 
12
- # Skills Catalog (90 skills)
12
+ # Skills Catalog
13
13
 
14
14
  All functionality is unified under skills. Task and hybrid skills are user-invocable as slash commands. Knowledge skills provide domain patterns auto-loaded by agents.
15
15
 
package/llms-full.txt CHANGED
@@ -1,6 +1,6 @@
1
1
  # ai-toolkit
2
2
 
3
- > Professional-grade Claude Code toolkit: 90 skills, 44 agents, machine-enforced constitution, quality hooks.
3
+ > Professional-grade Claude Code toolkit: skills, agents, machine-enforced constitution, quality hooks.
4
4
 
5
5
  ## Documentation
6
6
 
@@ -12,16 +12,16 @@
12
12
  ## Knowledge Base
13
13
 
14
14
  - [Best Practices](kb/best-practices/README.md)
15
+ - [No Hardcoded Counts in Secondary Docs](kb/best-practices/no-hardcoded-counts.md)
15
16
  - [How-To Guides](kb/howto/README.md)
16
17
  - [SOP: Claude Toolkit Maintenance](kb/procedures/maintenance-sop.md)
17
- - [Agents Catalog (44 agents)](kb/reference/agents-catalog.md)
18
+ - [Agents Catalog](kb/reference/agents-catalog.md)
18
19
  - [Anti-Pattern Registry Format](kb/reference/anti-pattern-registry-format.md)
19
20
  - [AI Toolkit Architecture](kb/reference/architecture-overview.md)
20
21
  - [Config Benchmark](kb/reference/benchmark-config.md)
21
22
  - [CI Integration](kb/reference/ci-integration.md)
22
23
  - [Claude Ecosystem Benchmark Snapshot](kb/reference/claude-ecosystem-benchmark-snapshot.md)
23
24
  - [Claude Ecosystem Expansion Foundations](kb/reference/claude-ecosystem-expansion-foundations.md)
24
- - [Commands Catalog (DEPRECATED)](kb/reference/commands-catalog.md)
25
25
  - [Plan: Competitive Features — ai-toolkit](kb/reference/competitive-features-implementation.md)
26
26
  - [Distribution Model](kb/reference/distribution-model.md)
27
27
  - [Extension API Reference](kb/reference/extension-api.md)
@@ -37,7 +37,7 @@
37
37
  - [Plugin Pack Conventions](kb/reference/plugin-pack-conventions.md)
38
38
  - [Quick Wins Implementation Summary](kb/reference/quick-wins-implementation-summary.md)
39
39
  - [Skill Templates](kb/reference/skill-templates.md)
40
- - [Skills Catalog (90 skills)](kb/reference/skills-catalog.md)
40
+ - [Skills Catalog](kb/reference/skills-catalog.md)
41
41
  - [Skills Unification Model](kb/reference/skills-unification.md)
42
42
  - [Usage Statistics](kb/reference/stats.md)
43
43
  - [Config Sync](kb/reference/sync.md)
@@ -201,6 +201,52 @@ Guidelines and recommendations. Guides will be added here as they are created.
201
201
 
202
202
  ---
203
203
 
204
+ ## kb/best-practices/no-hardcoded-counts.md
205
+
206
+ ---
207
+ title: "No Hardcoded Counts in Secondary Docs"
208
+ category: best-practices
209
+ service: ai-toolkit
210
+ tags: [counts, documentation, maintenance, skills, agents, hooks, tests]
211
+ created: "2026-04-07"
212
+ last_updated: "2026-04-07"
213
+ description: "Counts (skills, agents, hooks, tests) should only appear in README.md and manifest.json. All other docs must NOT contain hardcoded numbers to avoid drift."
214
+ ---
215
+
216
+ # No Hardcoded Counts in Secondary Docs
217
+
218
+ ## Rule
219
+
220
+ Hardcoded counts (skills, agents, hooks, tests, plugins) are allowed ONLY in:
221
+ - **README.md** — badges, "What You Get" table, comparison table
222
+ - **manifest.json** — module descriptions
223
+ - **package.json** — description field
224
+
225
+ All other files (CLAUDE.md, ARCHITECTURE.md, KB docs, plugin.json, llms.txt, AGENTS.md, copilot-instructions, rules, GEMINI.md) must NOT contain hardcoded counts like "90 skills" or "44 agents".
226
+
227
+ ## Why
228
+
229
+ Every time a skill, agent, or hook is added/removed, dozens of files need updating. This causes constant drift and stale counts that erode trust. Consolidating to 2-3 files makes maintenance tractable.
230
+
231
+ ## How to Apply
232
+
233
+ - In secondary docs, use relative language: "all agents", "the full skill set", "available hooks"
234
+ - If a doc MUST reference scale, use: "see README.md for current counts"
235
+ - `validate.py` checks counts only in README.md badges — that's sufficient
236
+ - When adding skills/agents/hooks, update ONLY: README.md badges + manifest.json descriptions
237
+
238
+ ## Anti-Pattern
239
+
240
+ ```markdown
241
+ # BAD — hardcoded count in ARCHITECTURE.md
242
+ Shared AI development toolkit — 90 skills, 44 agents
243
+
244
+ # GOOD — no count
245
+ Shared AI development toolkit with multi-platform support
246
+ ```
247
+
248
+ ---
249
+
204
250
  ## kb/howto/README.md
205
251
 
206
252
  ---
@@ -535,13 +581,13 @@ title: "AI Toolkit - Agents Catalog"
535
581
  category: reference
536
582
  service: ai-toolkit
537
583
  tags: [agents, catalog, roles, ai-development]
538
- version: "2.0.0"
584
+ version: "1.3.10"
539
585
  created: "2026-03-23"
540
586
  last_updated: "2026-03-25"
541
- description: "Complete catalog of 44 specialized agents with roles, models, and use cases."
587
+ description: "Complete catalog of specialized agents with roles, models, and use cases."
542
588
  ---
543
589
 
544
- # Agents Catalog (44 agents)
590
+ # Agents Catalog
545
591
 
546
592
  ## By Category
547
593
 
@@ -887,7 +933,7 @@ title: "AI Toolkit - Architecture Overview"
887
933
  category: reference
888
934
  service: ai-toolkit
889
935
  tags: [architecture, overview, design, structure]
890
- version: "1.0.0"
936
+ version: "1.3.10"
891
937
  created: "2026-03-23"
892
938
  last_updated: "2026-04-07"
893
939
  description: "Architecture of ai-toolkit: directory layout, global install model, skill tiers, and integration with projects."
@@ -897,7 +943,7 @@ description: "Architecture of ai-toolkit: directory layout, global install model
897
943
 
898
944
  ## Purpose
899
945
 
900
- Shared, project-agnostic AI development toolkit for Claude Code (and compatible assistants like Cursor, Windsurf, Copilot, Gemini, Cline, Roo Code, Aider, and Augment). Provides 44 specialized agents, 90 skills (slash commands + knowledge), expanded lifecycle hooks, persona presets, and experimental opt-in plugin packs that teams can adopt separately from the default global install.
946
+ Shared, project-agnostic AI development toolkit for Claude Code (and compatible assistants like Cursor, Windsurf, Copilot, Gemini, Cline, Roo Code, Aider, and Augment). Provides specialized agents, skills (slash commands + knowledge), expanded lifecycle hooks, persona presets, and experimental opt-in plugin packs that teams can adopt separately from the default global install.
901
947
 
902
948
  ## Design Principles
903
949
 
@@ -914,8 +960,8 @@ ai-toolkit/
914
960
  bin/
915
961
  ai-toolkit.js # CLI entry point (install, init, add-rule, ...)
916
962
  app/ # All toolkit components
917
- agents/ # 44 agent definitions (.md + YAML frontmatter)
918
- skills/ # 90 skills: task, hybrid, knowledge
963
+ agents/ # Agent definitions (.md + YAML frontmatter)
964
+ skills/ # skills: task, hybrid, knowledge
919
965
  rules/ # Rules auto-injected into ~/.claude/CLAUDE.md
920
966
  hooks/ # Hook scripts (copied to ~/.ai-toolkit/hooks/)
921
967
  hooks.json # Hook definitions (merged into ~/.claude/settings.json)
@@ -1493,32 +1539,6 @@ The reference benchmark set is intentionally curated:
1493
1539
 
1494
1540
  ---
1495
1541
 
1496
- ## kb/reference/commands-catalog.md
1497
-
1498
- ---
1499
- title: "AI Toolkit - Commands Catalog (DEPRECATED)"
1500
- category: reference
1501
- service: ai-toolkit
1502
- tags: [commands, deprecated]
1503
- version: "1.0.0"
1504
- created: "2026-03-23"
1505
- last_updated: "2026-03-28"
1506
- description: "DEPRECATED: All slash commands are implemented as skills. See skills-catalog.md for the current catalog."
1507
- ---
1508
-
1509
- # Commands Catalog (DEPRECATED)
1510
-
1511
- All slash commands have been migrated to skills.
1512
-
1513
- See **[Skills Catalog](skills-catalog.md)** for the complete list of 90 skills, including:
1514
- - **28 Task Skills** — formerly standalone commands and creator workflows (e.g., `/commit`, `/test`, `/deploy`, `/hook-creator`, `/plugin-creator`)
1515
- - **30 Hybrid Skills** — slash commands that also provide agent knowledge (e.g., `/review`, `/debug`, `/plan`, `/tdd`, `/write-a-prd`, `/council`, `/introspect`)
1516
- - **32 Knowledge Skills** — domain patterns auto-loaded by agents (e.g., `brand-voice`, `clean-code`, `testing-patterns`)
1517
-
1518
- Slash command syntax (`/command`) continues to work. The underlying implementation moved from `app/commands/` to `app/skills/`.
1519
-
1520
- ---
1521
-
1522
1542
  ## kb/reference/competitive-features-implementation.md
1523
1543
 
1524
1544
  ---
@@ -1558,7 +1578,7 @@ Strengthen ai-toolkit's competitive position by implementing 10 features from co
1558
1578
  **Key design principle:** ai-toolkit is a **generic toolkit** — it does NOT know about rag-mcp or any specific consumer. Consumers (like rag-mcp) use ai-toolkit's public API (`inject-rule`, `inject-hook`, `merge-hooks`) to add their own rules and hooks.
1559
1579
 
1560
1580
  **State before plan:** 88 skills, 47 agents, 14 hooks, 9 editor integrations
1561
- **State after plan:** 90 skills, 44 agents, 21 hooks, 70 language rules, 25 MCP templates, extension API
1581
+ **State after plan:** See manifest.json for current counts. Target: skills, agents, hooks, 70 language rules, 25 MCP templates, extension API
1562
1582
 
1563
1583
  ---
1564
1584
 
@@ -1907,7 +1927,7 @@ All entries are tagged with "_source": "<source-name>" for idempotent updates.
1907
1927
  "required": true
1908
1928
  },
1909
1929
  "agents": {
1910
- "description": "44 specialized agents",
1930
+ "description": "Specialized agents",
1911
1931
  "files": ["app/agents/*.md"],
1912
1932
  "required": false,
1913
1933
  "default": true
@@ -3454,8 +3474,8 @@ Modules are defined in `manifest.json` at the repository root. There are 17 modu
3454
3474
  | Module | Description | In Profile |
3455
3475
  |--------|-------------|-----------|
3456
3476
  | `core` | Core hooks and essential skills | minimal, standard, strict, full |
3457
- | `agents` | 44 specialized agents | standard, strict, full |
3458
- | `skills` | 90 skills (task, hybrid, knowledge) | standard, strict, full |
3477
+ | `agents` | Specialized agents | standard, strict, full |
3478
+ | `skills` | All skills (task, hybrid, knowledge) | standard, strict, full |
3459
3479
  | `rules-common` | Common coding rules (5 files) | standard, strict, full |
3460
3480
  | `rules-typescript` | TypeScript-specific rules (5 files) | auto-detect |
3461
3481
  | `rules-python` | Python-specific rules (5 files) | auto-detect |
@@ -4067,13 +4087,13 @@ title: "AI Toolkit - Skills Catalog"
4067
4087
  category: reference
4068
4088
  service: ai-toolkit
4069
4089
  tags: [skills, domain-knowledge, catalog, task-skills, hybrid-skills]
4070
- version: "1.3.8"
4090
+ version: "1.3.10"
4071
4091
  created: "2026-03-23"
4072
4092
  last_updated: "2026-04-07"
4073
- description: "Complete catalog of 90 skills: 28 task, 30 hybrid, 32 knowledge. Includes effort levels, skill-scoped hooks, executable scripts, security auditor, and persona presets."
4093
+ description: "Complete skills catalog with task, hybrid, and knowledge skills. Includes effort levels, skill-scoped hooks, executable scripts, security auditor, and persona presets."
4074
4094
  ---
4075
4095
 
4076
- # Skills Catalog (90 skills)
4096
+ # Skills Catalog
4077
4097
 
4078
4098
  All functionality is unified under skills. Task and hybrid skills are user-invocable as slash commands. Knowledge skills provide domain patterns auto-loaded by agents.
4079
4099
 
package/llms.txt CHANGED
@@ -1,6 +1,6 @@
1
1
  # ai-toolkit
2
2
 
3
- > Professional-grade Claude Code toolkit: 90 skills, 44 agents, machine-enforced constitution, quality hooks.
3
+ > Professional-grade Claude Code toolkit: skills, agents, machine-enforced constitution, quality hooks.
4
4
 
5
5
  ## Documentation
6
6
 
@@ -12,16 +12,16 @@
12
12
  ## Knowledge Base
13
13
 
14
14
  - [Best Practices](kb/best-practices/README.md)
15
+ - [No Hardcoded Counts in Secondary Docs](kb/best-practices/no-hardcoded-counts.md)
15
16
  - [How-To Guides](kb/howto/README.md)
16
17
  - [SOP: Claude Toolkit Maintenance](kb/procedures/maintenance-sop.md)
17
- - [Agents Catalog (44 agents)](kb/reference/agents-catalog.md)
18
+ - [Agents Catalog](kb/reference/agents-catalog.md)
18
19
  - [Anti-Pattern Registry Format](kb/reference/anti-pattern-registry-format.md)
19
20
  - [AI Toolkit Architecture](kb/reference/architecture-overview.md)
20
21
  - [Config Benchmark](kb/reference/benchmark-config.md)
21
22
  - [CI Integration](kb/reference/ci-integration.md)
22
23
  - [Claude Ecosystem Benchmark Snapshot](kb/reference/claude-ecosystem-benchmark-snapshot.md)
23
24
  - [Claude Ecosystem Expansion Foundations](kb/reference/claude-ecosystem-expansion-foundations.md)
24
- - [Commands Catalog (DEPRECATED)](kb/reference/commands-catalog.md)
25
25
  - [Plan: Competitive Features — ai-toolkit](kb/reference/competitive-features-implementation.md)
26
26
  - [Distribution Model](kb/reference/distribution-model.md)
27
27
  - [Extension API Reference](kb/reference/extension-api.md)
@@ -37,7 +37,7 @@
37
37
  - [Plugin Pack Conventions](kb/reference/plugin-pack-conventions.md)
38
38
  - [Quick Wins Implementation Summary](kb/reference/quick-wins-implementation-summary.md)
39
39
  - [Skill Templates](kb/reference/skill-templates.md)
40
- - [Skills Catalog (90 skills)](kb/reference/skills-catalog.md)
40
+ - [Skills Catalog](kb/reference/skills-catalog.md)
41
41
  - [Skills Unification Model](kb/reference/skills-unification.md)
42
42
  - [Usage Statistics](kb/reference/stats.md)
43
43
  - [Config Sync](kb/reference/sync.md)
package/manifest.json CHANGED
@@ -1,5 +1,5 @@
1
1
  {
2
- "version": "1.0.0",
2
+ "version": "1.3.10",
3
3
  "components": {
4
4
  "agents": {
5
5
  "description": "44 specialized agents (orchestrator, backend, frontend, security, devops, etc.)",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@softspark/ai-toolkit",
3
- "version": "1.3.8",
3
+ "version": "1.3.10",
4
4
  "description": "Professional-grade AI coding toolkit: 90 skills, 44 agents, multi-platform support (Claude, Cursor, Windsurf, Copilot, Gemini, Cline, Roo Code, Aider, Augment), machine-enforced safety constitution, persona presets, skill security auditor, expanded lifecycle hooks, 11 plugin packs, and benchmark tooling.",
5
5
  "keywords": [
6
6
  "claude",
package/scripts/generate_augment.py CHANGED
@@ -31,7 +31,7 @@ if __name__ == "__main__":
31
31
  ],
32
32
  "intro_template": (
33
33
  "This repository uses the ai-toolkit — a shared AI development toolkit"
34
- " with {agents} specialized agent personas and {skills} skills."
34
+ " with specialized agent personas and skills."
35
35
  ),
36
36
  "agents_section": "## Available Agent Personas",
37
37
  "agents_intro": "Apply the expertise of these agents when working on relevant tasks:",
package/scripts/generate_cline.py CHANGED
@@ -20,7 +20,7 @@ if __name__ == "__main__":
20
20
  ],
21
21
  "intro_template": (
22
22
  "This repository uses the ai-toolkit — a shared AI development toolkit"
23
- " with {agents} specialized agent personas and {skills} skills."
23
+ " with specialized agent personas and skills."
24
24
  ),
25
25
  "agents_section": "## Available Agent Personas",
26
26
  "agents_intro": "Apply the expertise of these agents when working on relevant tasks:",
package/scripts/generate_copilot.py CHANGED
@@ -16,7 +16,7 @@ if __name__ == "__main__":
16
16
  "title": "# GitHub Copilot Instructions",
17
17
  "intro_template": (
18
18
  "This repository uses the ai-toolkit — a shared AI development toolkit"
19
- " with {agents} specialized agent personas and {skills} skills."
19
+ " with specialized agent personas and skills."
20
20
  ),
21
21
  "agents_section": "## Available Agent Personas",
22
22
  "agents_intro": "Apply the expertise of these agents when working on relevant tasks:",
package/scripts/generate_cursor_rules.py CHANGED
@@ -20,7 +20,7 @@ if __name__ == "__main__":
20
20
  ],
21
21
  "intro_template": (
22
22
  "This repository uses the ai-toolkit — a shared AI development toolkit"
23
- " with {agents} specialized agent personas and {skills} skills."
23
+ " with specialized agent personas and skills."
24
24
  ),
25
25
  "agents_section": "## Available Agent Personas",
26
26
  "agents_intro": "Apply the expertise of these agents when working on relevant tasks:",
package/scripts/generate_gemini.py CHANGED
@@ -15,13 +15,13 @@ if __name__ == "__main__":
15
15
  render_generator({
16
16
  "title": "# AI Toolkit — Gemini CLI Configuration",
17
17
  "intro_template": (
18
- "Shared AI development toolkit with {agents} specialized agents,"
19
- " {skills} skills, quality hooks, and a safety constitution."
18
+ "Shared AI development toolkit with specialized agents,"
19
+ " skills, quality hooks, and a safety constitution."
20
20
  ),
21
- "agents_section": "## Available Agents ({agents})",
21
+ "agents_section": "## Available Agents",
22
22
  "agents_intro": "Specialized agent personas — apply their expertise for relevant tasks:",
23
23
  "agents_format": "bullets",
24
- "skills_section": "## Available Skills ({skills})",
24
+ "skills_section": "## Available Skills",
25
25
  "skills_intro": "Skills are invocable slash commands or auto-loaded knowledge sources:",
26
26
  "skills_format": "bullets",
27
27
  "guidelines": ["quality_standards", "workflow"],
package/scripts/generate_llms_txt.py CHANGED
@@ -65,8 +65,8 @@ def generate_index() -> None:
65
65
  print("# ai-toolkit")
66
66
  print()
67
67
  print(
68
- f"> Professional-grade Claude Code toolkit: {skills} skills,"
69
- f" {agents} agents, machine-enforced constitution, quality hooks."
68
+ "> Professional-grade Claude Code toolkit: skills,"
69
+ " agents, machine-enforced constitution, quality hooks."
70
70
  )
71
71
  print()
72
72
  print("## Documentation")
package/scripts/generate_windsurf.py CHANGED
@@ -20,7 +20,7 @@ if __name__ == "__main__":
20
20
  ],
21
21
  "intro_template": (
22
22
  "This repository uses the ai-toolkit — a shared AI development toolkit"
23
- " with {agents} specialized agent personas and {skills} skills."
23
+ " with specialized agent personas and skills."
24
24
  ),
25
25
  "agents_section": "## Available Agent Personas",
26
26
  "agents_intro": "Apply the expertise of these agents when working on relevant tasks:",
package/scripts/generator_base.py CHANGED
@@ -12,7 +12,7 @@ Usage::
12
12
  render_generator({
13
13
  "title": "# Cursor Rules",
14
14
  "header_lines": ["# Auto-generated..."],
15
- "intro_template": "This repo uses ai-toolkit with {agents} agents and {skills} skills.",
15
+ "intro_template": "This repo uses ai-toolkit with specialized agents and skills.",
16
16
  "agents_section": "## Available Agent Personas",
17
17
  "agents_format": "headings", # "headings" or "bullets"
18
18
  "agents_level": "##", # heading level (for "headings" format)
@@ -57,7 +57,7 @@ def render_generator(config: dict) -> None:
57
57
  Config keys:
58
58
  title: str — main heading line
59
59
  header_lines: list[str] — additional header lines (comments etc)
60
- intro_template: str — format string with {agents} and {skills}
60
+ intro_template: str — intro text (may use {agents}/{skills} placeholders)
61
61
  agents_section: str — heading for agents section
62
62
  agents_format: "headings" | "bullets"
63
63
  agents_level: str — heading level for headings format (default "##")
package/scripts/install_steps/install_state.py CHANGED
@@ -88,6 +88,11 @@ def record_install(
88
88
 
89
89
  save_state(state)
90
90
 
91
+ # Clear version check cache (version may have changed)
92
+ cache_file = _state_path().parent / "version-check.json"
93
+ if cache_file.is_file():
94
+ cache_file.unlink()
95
+
91
96
 
92
97
  def print_status() -> None:
93
98
  """Print a human-readable summary of the install state."""
package/kb/reference/commands-catalog.md DELETED
@@ -1,21 +0,0 @@
1
- ---
2
- title: "AI Toolkit - Commands Catalog (DEPRECATED)"
3
- category: reference
4
- service: ai-toolkit
5
- tags: [commands, deprecated]
6
- version: "1.0.0"
7
- created: "2026-03-23"
8
- last_updated: "2026-03-28"
9
- description: "DEPRECATED: All slash commands are implemented as skills. See skills-catalog.md for the current catalog."
10
- ---
11
-
12
- # Commands Catalog (DEPRECATED)
13
-
14
- All slash commands have been migrated to skills.
15
-
16
- See **[Skills Catalog](skills-catalog.md)** for the complete list of 90 skills, including:
17
- - **28 Task Skills** — formerly standalone commands and creator workflows (e.g., `/commit`, `/test`, `/deploy`, `/hook-creator`, `/plugin-creator`)
18
- - **30 Hybrid Skills** — slash commands that also provide agent knowledge (e.g., `/review`, `/debug`, `/plan`, `/tdd`, `/write-a-prd`, `/council`, `/introspect`)
19
- - **32 Knowledge Skills** — domain patterns auto-loaded by agents (e.g., `brand-voice`, `clean-code`, `testing-patterns`)
20
-
21
- Slash command syntax (`/command`) continues to work. The underlying implementation moved from `app/commands/` to `app/skills/`.