aiwg 2026.6.1 → 2026.6.2

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 (120) hide show
  1. package/agentic/code/addons/aiwg-utils/rules/delivery-policy.md +32 -0
  2. package/agentic/code/addons/aiwg-utils/skills/aiwg-utils-quickref/SKILL.md +11 -0
  3. package/agentic/code/addons/aiwg-utils/skills/new-project/SKILL.md +8 -1
  4. package/agentic/code/addons/aiwg-utils/skills/scaffold-addon/SKILL.md +8 -2
  5. package/agentic/code/addons/aiwg-utils/skills/scaffold-extension/SKILL.md +8 -1
  6. package/agentic/code/addons/aiwg-utils/skills/scaffold-framework/SKILL.md +8 -2
  7. package/agentic/code/addons/aiwg-utils/skills/soul-create/SKILL.md +8 -1
  8. package/agentic/code/addons/aiwg-utils/skills/steward/SKILL.md +20 -0
  9. package/agentic/code/addons/aiwg-utils/skills/steward-quickref/SKILL.md +87 -0
  10. package/agentic/code/agents/personas/aiwg-finder.md +6 -0
  11. package/agentic/code/agents/personas/aiwg-orchestrator.md +8 -1
  12. package/agentic/code/agents/personas/aiwg-reviewer.md +8 -1
  13. package/agentic/code/agents/personas/aiwg-security.md +8 -1
  14. package/agentic/code/agents/personas/aiwg-steward.md +12 -1
  15. package/agentic/code/agents/personas/aiwg-writer.md +8 -1
  16. package/agentic/code/agents/personas/mc-conductor.md +7 -1
  17. package/agentic/code/extensions/research-brainstorming/README.md +112 -0
  18. package/agentic/code/extensions/research-brainstorming/fixtures/linked-retrieval.json +31 -0
  19. package/agentic/code/extensions/research-brainstorming/fixtures/loose-session.md +41 -0
  20. package/agentic/code/extensions/research-brainstorming/fixtures/promotion-ref-candidate-corpus-root.md +31 -0
  21. package/agentic/code/extensions/research-brainstorming/fixtures/promotion-research-task.md +25 -0
  22. package/agentic/code/extensions/research-brainstorming/manifest.json +57 -0
  23. package/agentic/code/extensions/research-brainstorming/skills/research-brainstorm/SKILL.md +46 -0
  24. package/agentic/code/extensions/research-brainstorming/skills/research-brainstorm-map/SKILL.md +34 -0
  25. package/agentic/code/extensions/research-brainstorming/skills/research-brainstorm-promote/SKILL.md +51 -0
  26. package/agentic/code/extensions/research-brainstorming/skills/research-brainstorm-query/SKILL.md +38 -0
  27. package/agentic/code/extensions/research-brainstorming/templates/brainstorm-note.md +37 -0
  28. package/agentic/code/extensions/research-brainstorming/templates/exploration-plan.md +27 -0
  29. package/agentic/code/extensions/research-brainstorming/templates/promotion-record.md +26 -0
  30. package/agentic/code/frameworks/research-complete/manifest.json +14 -0
  31. package/agentic/code/frameworks/sdlc-complete/templates/codex/config.toml.aiwg-template +2 -1
  32. package/agentic/code/frameworks/sdlc-complete/templates/cursor/cursor-mcp-full.json +2 -1
  33. package/agentic/code/frameworks/sdlc-complete/templates/cursor/cursor-mcp-minimal.json +2 -1
  34. package/agentic/code/frameworks/sdlc-complete/templates/hermes/AGENTS.md.aiwg-template +1 -1
  35. package/agentic/code/frameworks/sdlc-complete/templates/hermes/hermes-config-minimal.yaml +3 -2
  36. package/agentic/code/frameworks/sdlc-complete/templates/hermes/hermes-config-prompts.yaml +1 -1
  37. package/agentic/code/frameworks/sdlc-complete/templates/hermes/skills/aiwg-orchestrate/SKILL.md +1 -1
  38. package/agentic/code/frameworks/sdlc-complete/templates/warp/warp-mcp-full.json +2 -1
  39. package/agentic/code/frameworks/sdlc-complete/templates/warp/warp-mcp-minimal.json +2 -1
  40. package/agentic/code/frameworks/sdlc-complete/templates/windsurf/windsurf-mcp-full.json +2 -1
  41. package/agentic/code/frameworks/sdlc-complete/templates/windsurf/windsurf-mcp-minimal.json +2 -1
  42. package/dist/src/agents/agent-validator.d.ts.map +1 -1
  43. package/dist/src/agents/agent-validator.js +24 -0
  44. package/dist/src/agents/agent-validator.js.map +1 -1
  45. package/dist/src/agents/types.d.ts +7 -0
  46. package/dist/src/agents/types.d.ts.map +1 -1
  47. package/dist/src/artifacts/cli.d.ts.map +1 -1
  48. package/dist/src/artifacts/cli.js +28 -1
  49. package/dist/src/artifacts/cli.js.map +1 -1
  50. package/dist/src/artifacts/discover-facets.d.ts +95 -0
  51. package/dist/src/artifacts/discover-facets.d.ts.map +1 -0
  52. package/dist/src/artifacts/discover-facets.js +286 -0
  53. package/dist/src/artifacts/discover-facets.js.map +1 -0
  54. package/dist/src/artifacts/index-status.d.ts +72 -0
  55. package/dist/src/artifacts/index-status.d.ts.map +1 -0
  56. package/dist/src/artifacts/index-status.js +168 -0
  57. package/dist/src/artifacts/index-status.js.map +1 -0
  58. package/dist/src/artifacts/query-engine.d.ts.map +1 -1
  59. package/dist/src/artifacts/query-engine.js +65 -7
  60. package/dist/src/artifacts/query-engine.js.map +1 -1
  61. package/dist/src/artifacts/types.d.ts +19 -2
  62. package/dist/src/artifacts/types.d.ts.map +1 -1
  63. package/dist/src/artifacts/types.js +35 -6
  64. package/dist/src/artifacts/types.js.map +1 -1
  65. package/dist/src/cli/command-log.d.ts +51 -0
  66. package/dist/src/cli/command-log.d.ts.map +1 -0
  67. package/dist/src/cli/command-log.js +271 -0
  68. package/dist/src/cli/command-log.js.map +1 -0
  69. package/dist/src/cli/handlers/cockpit.d.ts +12 -0
  70. package/dist/src/cli/handlers/cockpit.d.ts.map +1 -0
  71. package/dist/src/cli/handlers/cockpit.js +137 -0
  72. package/dist/src/cli/handlers/cockpit.js.map +1 -0
  73. package/dist/src/cli/handlers/command-log.d.ts +3 -0
  74. package/dist/src/cli/handlers/command-log.d.ts.map +1 -0
  75. package/dist/src/cli/handlers/command-log.js +37 -0
  76. package/dist/src/cli/handlers/command-log.js.map +1 -0
  77. package/dist/src/cli/handlers/help.js +4 -0
  78. package/dist/src/cli/handlers/help.js.map +1 -1
  79. package/dist/src/cli/handlers/index.d.ts +3 -1
  80. package/dist/src/cli/handlers/index.d.ts.map +1 -1
  81. package/dist/src/cli/handlers/index.js +7 -2
  82. package/dist/src/cli/handlers/index.js.map +1 -1
  83. package/dist/src/cli/handlers/subcommands.d.ts.map +1 -1
  84. package/dist/src/cli/handlers/subcommands.js +58 -6
  85. package/dist/src/cli/handlers/subcommands.js.map +1 -1
  86. package/dist/src/cli/handlers/use.d.ts.map +1 -1
  87. package/dist/src/cli/handlers/use.js +4 -0
  88. package/dist/src/cli/handlers/use.js.map +1 -1
  89. package/dist/src/cli/provider-deployment-plan.d.ts +38 -0
  90. package/dist/src/cli/provider-deployment-plan.d.ts.map +1 -0
  91. package/dist/src/cli/provider-deployment-plan.js +126 -0
  92. package/dist/src/cli/provider-deployment-plan.js.map +1 -0
  93. package/dist/src/cli/router.d.ts.map +1 -1
  94. package/dist/src/cli/router.js +54 -3
  95. package/dist/src/cli/router.js.map +1 -1
  96. package/dist/src/config/aiwg-config.d.ts +47 -0
  97. package/dist/src/config/aiwg-config.d.ts.map +1 -1
  98. package/dist/src/config/aiwg-config.js +3 -0
  99. package/dist/src/config/aiwg-config.js.map +1 -1
  100. package/dist/src/config/cli.js +26 -0
  101. package/dist/src/config/cli.js.map +1 -1
  102. package/dist/src/extensions/commands/definitions.d.ts +2 -0
  103. package/dist/src/extensions/commands/definitions.d.ts.map +1 -1
  104. package/dist/src/extensions/commands/definitions.js +68 -0
  105. package/dist/src/extensions/commands/definitions.js.map +1 -1
  106. package/dist/src/extensions/validation.d.ts +3 -3
  107. package/dist/src/extensions/validation.d.ts.map +1 -1
  108. package/dist/src/extensions/validation.js +11 -1
  109. package/dist/src/extensions/validation.js.map +1 -1
  110. package/dist/src/mcp/cli.mjs +7 -5
  111. package/dist/src/mcp/server.mjs +2 -395
  112. package/dist/src/mcp/test-simple.mjs +7 -149
  113. package/dist/src/mcp/tools/orchestration.mjs +289 -0
  114. package/dist/src/mcp/tools/subsystems.mjs +4 -1
  115. package/package.json +5 -2
  116. package/tools/agents/deploy-agents.mjs +73 -0
  117. package/tools/agents/providers/base.mjs +282 -18
  118. package/tools/cli/doctor.mjs +138 -14
  119. package/tools/lint/agent-def-sizes.mjs +145 -0
  120. package/tools/release/cut-tag.sh +31 -17
@@ -34,8 +34,20 @@ Located at `.aiwg/aiwg.config` (top level), the `delivery` block contains:
34
34
  "delivery": {
35
35
  "mode": "direct" | "feature-branch" | "pr-required",
36
36
  "default_branch": "main",
37
+ "committer": {
38
+ "name": "Joseph Magly",
39
+ "email": "1159087+jmagly@users.noreply.github.com"
40
+ },
41
+ "signing": {
42
+ "format": "openpgp" | "ssh" | "x509",
43
+ "key": "0117DAAA677A5BF2",
44
+ "key_file": "~/.config/.../signing.key",
45
+ "program": "gpg",
46
+ "enforce": "commits" | "tags" | "all"
47
+ },
37
48
  "require_ci_green": true,
38
49
  "force_push_policy": "never" | "main-only-blocked" | "allowed",
50
+ "require_signed_commits": true,
39
51
  "auto_close_issues": true,
40
52
  "issue_comment_on_cycle": true,
41
53
  "rationale": "Free-text explaining why this mode was chosen"
@@ -44,6 +56,11 @@ Located at `.aiwg/aiwg.config` (top level), the `delivery` block contains:
44
56
  "primary": "origin",
45
57
  "issue_tracker": "origin",
46
58
  "ci": "origin",
59
+ "tracker_actor": {
60
+ "login": "roctinam",
61
+ "via": "tea" | "gh" | "mcp" | "api",
62
+ "forbid_actors": ["roctibot"]
63
+ },
47
64
  "secondary": [ ... ]
48
65
  }
49
66
  }
@@ -134,6 +151,19 @@ use the AIWG issue CLI for local issue operations. Do not invent a parallel
134
151
  issue-provider setting when `remotes.issue_tracker` already answers where
135
152
  issues live.
136
153
 
154
+ ### Rule 3.5: Use Configured Delivery Identity For Mutations
155
+
156
+ When delivery identity fields are present, they are authoritative for writes:
157
+
158
+ - Commits → use `delivery.committer.name` / `delivery.committer.email` when set; otherwise inherit normal `git config`.
159
+ - Commit and tag signing → if `delivery.require_signed_commits: true` or `delivery.signing.enforce` covers the artifact being written, use `delivery.signing.format`, `delivery.signing.key` or `delivery.signing.key_file`, and `delivery.signing.program` where applicable.
160
+ - Issue comments, issue closures, PRs, labels, and milestone writes → use `remotes.tracker_actor.login` through `remotes.tracker_actor.via` when set.
161
+ - Never author delivery mutations as any login listed in `remotes.tracker_actor.forbid_actors`.
162
+
163
+ Read-only tracker operations may use any available credential, including an MCP token, because they do not create attribution. Delivery mutations must use the configured actor or stop and ask the operator to configure/authorize a route.
164
+
165
+ `aiwg doctor` warns when signed commits are required but no signing key/key file is declared, and when an issue tracker remote is configured without a `tracker_actor` in a repo likely to perform tracker writes.
166
+
137
167
  ### Rule 4: Surface, Don't Re-Ask
138
168
 
139
169
  When the policy is already declared, do NOT use `AskUserQuestion` (or equivalent interactive prompt) to ask the user "feature branch or direct to main?" — the answer is already in the config. Instead:
@@ -178,5 +208,7 @@ This rule complements:
178
208
 
179
209
  - `aiwg config get --project delivery.mode` — CLI command to inspect current policy
180
210
  - `aiwg config set --project delivery.mode <mode>` — change project policy
211
+ - `aiwg config set --project delivery.signing.key <key-id>` — declare the signing key used for commits/tags
212
+ - `aiwg config set --project remotes.tracker_actor.login <login>` — declare the forge actor for issue/PR writes
181
213
  - `agentic/code/frameworks/sdlc-complete/skills/issue-create/SKILL.md` — example of a skill that reads `aiwg.config` properly
182
214
  - `agentic/code/frameworks/sdlc-complete/skills/flow-delivery-track/SKILL.md` — delivery workflow integration
@@ -191,6 +191,17 @@ aiwg discover "choose issue tracking backend" # → issue-workflow-guide
191
191
  aiwg discover "project-local customization" # → project-local quickstart / customization docs
192
192
  ```
193
193
 
194
+ ### Feature domains the steward owns (expansion / persona / project)
195
+
196
+ Three cross-cutting domains that no framework quickref owns — consult the `steward-quickref` kernel skill, or discover directly:
197
+
198
+ ```bash
199
+ aiwg discover "author an expansion" # → scaffold-extension / -addon / -framework
200
+ aiwg discover "create a persona" # → soul-create (author a SOUL identity)
201
+ aiwg discover "select a persona" # → persona agents (agentic/code/agents/personas/*)
202
+ aiwg discover "scaffold a project" # → new-project (+ new-bundle for project-local)
203
+ ```
204
+
194
205
  ### Project-local extensions (pilot → promote)
195
206
 
196
207
  AIWG supports authoring rules/skills/agents per-project under `.aiwg/{extensions,addons,frameworks,plugins}/<name>/` — pilot here, promote upstream when mature.
@@ -2,7 +2,14 @@
2
2
  namespace: aiwg
3
3
  name: new-project
4
4
  platforms: [all]
5
- description: Scaffold a new AIWG-managed project with .aiwg/ directory structure, starter CLAUDE.md, and framework registry
5
+ description: Scaffold a new AIWG-managed project with .aiwg/ directory structure, starter CLAUDE.md, and framework registry. Use when creating or scaffolding a new project.
6
+ triggers:
7
+ - scaffold a project
8
+ - create a project
9
+ - new project
10
+ - aiwg new
11
+ - project scaffolding
12
+ - project-local bundle
6
13
  ---
7
14
 
8
15
  <!-- AIWG-SKILL-CALLOUT -->
@@ -2,8 +2,14 @@
2
2
  namespace: aiwg
3
3
  name: scaffold-addon
4
4
  platforms: [all]
5
- description: Create a complete addon package structure inside agentic/code/addons/
6
-
5
+ description: Create a complete addon package structure inside agentic/code/addons/. Use when authoring an expansion (addon) for AIWG.
6
+ triggers:
7
+ - author an addon
8
+ - create an addon
9
+ - build an expansion addon
10
+ - addon authoring
11
+ - scaffold addon
12
+ - new expansion
7
13
  ---
8
14
 
9
15
  # Scaffold Addon
@@ -2,7 +2,14 @@
2
2
  namespace: aiwg
3
3
  name: scaffold-extension
4
4
  platforms: [all]
5
- description: Create a new extension package inside an existing framework's extensions/ directory
5
+ description: Create a new extension package inside an existing framework's extensions/ directory. Use when authoring an expansion (extension) for AIWG.
6
+ triggers:
7
+ - author an expansion
8
+ - build an extension
9
+ - create an extension package
10
+ - expansion authoring
11
+ - scaffold extension
12
+ - new expansion
6
13
  ---
7
14
 
8
15
  # Scaffold Extension
@@ -2,8 +2,14 @@
2
2
  namespace: aiwg
3
3
  name: scaffold-framework
4
4
  platforms: [all]
5
- description: Create a complete framework package structure inside agentic/code/frameworks/
6
-
5
+ description: Create a complete framework package structure inside agentic/code/frameworks/. Use when authoring an expansion (framework) for AIWG.
6
+ triggers:
7
+ - author a framework
8
+ - create a framework
9
+ - build a framework package
10
+ - framework authoring
11
+ - scaffold framework
12
+ - new expansion
7
13
  ---
8
14
 
9
15
  # Scaffold Framework
@@ -2,7 +2,14 @@
2
2
  namespace: aiwg
3
3
  name: soul-create
4
4
  platforms: [all]
5
- description: Generate a SOUL.md identity file from source material, interactive prompts, or an existing voice profile
5
+ description: Generate a SOUL.md identity file from source material, interactive prompts, or an existing voice profile. Use when authoring a persona / soul identity.
6
+ triggers:
7
+ - create a persona
8
+ - author a soul
9
+ - make an identity profile
10
+ - new persona
11
+ - generate SOUL.md
12
+ - persona authoring
6
13
  ---
7
14
 
8
15
  # soul-create
@@ -88,6 +88,26 @@ Route these intents directly:
88
88
  Canonical docs: `docs/customization/project-local-quickstart.md`, `docs/project-local/overview.md`, and `docs/project-local/manifest-reference.md`. Mention that project-local artifacts are trusted repo code and should be reviewed before deploy.
89
89
 
90
90
 
91
+ ## Feature-Domain Routing (proactive)
92
+
93
+ Three cross-cutting feature domains fall outside the framework quickrefs and were historically undiscoverable (#1623). The steward owns routing for them via the `steward-quickref` kernel skill. Be **proactive** — these are easy to miss:
94
+
95
+ | Domain | Canonical discover phrase | Owning capability |
96
+ |---|---|---|
97
+ | **Expansion authoring** (extension/addon/framework) | `aiwg discover "author an expansion"` | `scaffold-extension` / `-addon` / `-framework` |
98
+ | **Persona / SOUL** (author **and** select) | `aiwg discover "create a persona"` · `aiwg discover "select a persona"` | `soul-create`, persona agents under `agentic/code/agents/personas/` |
99
+ | **Project creation** | `aiwg discover "scaffold a project"` | `new-project`, `new-bundle` |
100
+
101
+ Routing protocol:
102
+
103
+ 1. **Volunteer the affordance** (Norman signifier). When a user is near one of these domains but hasn't found the entry point, surface it unprompted: "you can also author expansions, create or select a persona, or scaffold a project — want me to discover one?"
104
+ 2. **Discover, don't dead-end.** Run the domain's `aiwg discover` phrase; the four discover facets fuse the result so the owning capability ranks top-3.
105
+ 3. **Re-query on low confidence.** If the first pass is weak, broaden the phrase or try an adjacent domain before concluding "not found" — the `skill-discovery` rule forbids decline-without-search.
106
+ 4. **Show the selection.** `aiwg show <type> <name>` for the chosen capability.
107
+
108
+ See `steward-quickref` for the full anchor tables.
109
+
110
+
91
111
  ## Capability Matrix Source
92
112
 
93
113
  The authoritative source is `agentic/code/providers/capability-matrix.yaml`. Key features tracked:
@@ -0,0 +1,87 @@
1
+ ---
2
+ namespace: aiwg
3
+ name: steward-quickref
4
+ platforms: [all]
5
+ kernel: true
6
+ description: AUTO-INVOKE for steward feature-domain routing — authoring an expansion (extension/addon/framework), creating or selecting a persona/SOUL identity, or scaffolding a project. Thin discover anchors for the three domains the framework quickrefs don't own.
7
+ triggers:
8
+ - author an expansion
9
+ - create a persona
10
+ - select a persona
11
+ - scaffold a project
12
+ - steward routing
13
+ - which aiwg feature do i use
14
+ - expansion persona or project
15
+ ---
16
+
17
+ # Steward Quickref
18
+
19
+ Thin routing anchor for the three cross-cutting **feature domains** that no
20
+ single framework quickref owns: **expansion authoring**, **persona/SOUL
21
+ identity**, and **project creation**. This file holds *guidance + anchoring
22
+ phrases only* — it never inlines skill or agent bodies. The capability index
23
+ does the heavy lifting: each row gives the `aiwg discover` phrase that ranks
24
+ the owning capability top-3 (the four discover facets fuse the result —
25
+ see `src/artifacts/discover-facets.ts`).
26
+
27
+ > **Always discover, then show.** These phrases are the entry points; fetch the
28
+ > selected artifact with `aiwg show <type> <name>`. Do not browse the
29
+ > filesystem (`skill-discovery` rule).
30
+
31
+ ## Domain 1 — Expansion authoring (extension / addon / framework)
32
+
33
+ "How do I *build* an AIWG expansion?" — extensions, addons, and frameworks.
34
+
35
+ | You want to… | Discover phrase | Owning capability |
36
+ |---|---|---|
37
+ | Author an extension | `aiwg discover "author an expansion"` | `scaffold-extension` |
38
+ | Create an addon | `aiwg discover "create an addon"` | `scaffold-addon` |
39
+ | Build a framework | `aiwg discover "scaffold a framework"` | `scaffold-framework` |
40
+ | Pilot then graduate a project-local bundle | `aiwg discover "project-local bundle"` | `new-bundle`, `promote` |
41
+
42
+ ## Domain 2 — Persona / SOUL identity (author **and** select)
43
+
44
+ "How do I *make* or *choose* an identity/voice?" Personas live at
45
+ `agentic/code/agents/personas/`; SOUL profiles are managed by the `soul-*`
46
+ skills.
47
+
48
+ | You want to… | Discover phrase | Owning capability |
49
+ |---|---|---|
50
+ | Create / author a persona (SOUL) | `aiwg discover "create a persona"` | `soul-create` |
51
+ | Select / switch a persona identity | `aiwg discover "select a persona"` | persona agents + `soul-apply` |
52
+ | See the persona roster | `aiwg discover "persona"` | `agentic/code/agents/personas/*` |
53
+ | Manage SOUL lifecycle (enable/disable/blend) | `aiwg discover "soul"` | `soul-enable`, `soul-disable`, `soul-blend` |
54
+
55
+ > Runtime persona *selection UX* (catalog-pick) is an open research area — see
56
+ > spike #1625; for now, discover the persona then activate via the SOUL skills.
57
+
58
+ ## Domain 3 — Project creation (`aiwg new` + project-local)
59
+
60
+ "How do I *start* an AIWG project?"
61
+
62
+ | You want to… | Discover phrase | Owning capability |
63
+ |---|---|---|
64
+ | Scaffold a new project | `aiwg discover "scaffold a project"` | `new-project` |
65
+ | Create a project-local bundle | `aiwg discover "project-local bundle"` | `new-bundle` |
66
+
67
+ ## Domain 4 — Provider capability (native vs emulated)
68
+
69
+ "Does my provider support X natively, or does AIWG emulate it?" Route through
70
+ the steward, which reads `capability-matrix.yaml`.
71
+
72
+ | You want to… | Discover phrase / route |
73
+ |---|---|
74
+ | Check provider support for a capability | `aiwg discover "provider capability"` → `steward` |
75
+
76
+ ## Steward routing protocol
77
+
78
+ When a request maps to one of these domains:
79
+
80
+ 1. **Volunteer the affordance** (Norman signifier): if a user is working near a
81
+ domain but hasn't found it, surface it — "you can also author expansions,
82
+ create/select a persona, or scaffold a project; want me to discover one?"
83
+ 2. **Discover, don't dead-end.** Run the domain's `aiwg discover` phrase.
84
+ 3. **Re-query on low confidence.** If the first pass is weak, broaden the
85
+ phrase or try an adjacent domain phrase before concluding "not found" — the
86
+ `skill-discovery` rule forbids decline-without-search.
87
+ 4. **Show the selection.** `aiwg show <type> <name>` for the chosen capability.
@@ -1,6 +1,12 @@
1
1
  ---
2
2
  name: aiwg-finder
3
3
  description: Capability discovery and tool-selection specialist — the finder for AIWG's 400+ skills, agents, commands, and rules. Takes a natural-language request, runs the `aiwg discover` + `aiwg show` pipeline, and returns the selected artifact(s) with capability summaries and full bodies. Companion to aiwg-steward.
4
+ triggers:
5
+ - finder persona
6
+ - discovery persona
7
+ - capability finder identity
8
+ - select a finder persona
9
+ - persona
4
10
  model: claude-sonnet-4-6
5
11
  tools:
6
12
  - Bash
@@ -1,6 +1,13 @@
1
1
  ---
2
2
  name: aiwg-orchestrator
3
- description: Full SDLC orchestration persona for end-to-end project management
3
+ description: Full SDLC orchestration persona/identity for end-to-end project management. Use when selecting an orchestration persona to drive multi-agent SDLC workflows.
4
+ triggers:
5
+ - orchestrator persona
6
+ - sdlc orchestration identity
7
+ - project management persona
8
+ - select an orchestration persona
9
+ - persona
10
+ - soul
4
11
  model: claude-opus-4-7
5
12
  tools:
6
13
  - Task
@@ -1,6 +1,13 @@
1
1
  ---
2
2
  name: aiwg-reviewer
3
- description: Code review persona focused on quality, security, and best practices
3
+ description: Code review persona/identity focused on quality, security, and best practices. Use when selecting a reviewer persona for code or artifact review.
4
+ triggers:
5
+ - reviewer persona
6
+ - code review identity
7
+ - review voice
8
+ - select a review persona
9
+ - persona
10
+ - soul
4
11
  model: claude-sonnet-4-6
5
12
  tools:
6
13
  - Read
@@ -1,6 +1,13 @@
1
1
  ---
2
2
  name: aiwg-security
3
- description: Security audit persona for threat modeling and vulnerability assessment
3
+ description: Security audit persona/identity for threat modeling and vulnerability assessment. Use when selecting a security persona for review or threat work.
4
+ triggers:
5
+ - security persona
6
+ - security reviewer identity
7
+ - threat-model persona
8
+ - select a security persona
9
+ - persona
10
+ - soul
4
11
  model: claude-sonnet-4-6
5
12
  tools:
6
13
  - Read
@@ -1,6 +1,13 @@
1
1
  ---
2
2
  name: aiwg-steward
3
- description: Self-maintenance agent that uses AIWG CLI to keep the installation healthy, current, and correctly configured. Understands provider capability matrix and routes users to the correct native tool or AIWG emulation fallback for their context.
3
+ description: Self-maintenance steward persona/identity that uses the AIWG CLI to keep the installation healthy, current, and correctly configured. Understands the provider capability matrix and routes users to the correct native tool or AIWG emulation fallback for their context.
4
+ triggers:
5
+ - steward persona
6
+ - maintenance persona
7
+ - routing identity
8
+ - self-maintenance persona
9
+ - select a steward persona
10
+ - persona
4
11
  model: claude-sonnet-4-6
5
12
  tools:
6
13
  - Bash
@@ -65,6 +72,10 @@ Route these intents directly:
65
72
 
66
73
  Canonical docs: `docs/customization/project-local-quickstart.md`, `docs/project-local/overview.md`, and `docs/project-local/manifest-reference.md`. Mention that project-local artifacts are trusted repo code and should be reviewed before deploy.
67
74
 
75
+ ## Feature-Domain Routing (proactive)
76
+
77
+ Three cross-cutting domains were historically undiscoverable (#1623): **expansion authoring**, **persona/SOUL** (author and select), and **project creation**. Consult the `steward-quickref` kernel skill and route via `aiwg discover` — `"author an expansion"` → `scaffold-*`, `"create a persona"` / `"select a persona"` → `soul-create` + `agentic/code/agents/personas/*`, `"scaffold a project"` → `new-project`. Be proactive: volunteer the affordance when a user is near a domain but hasn't found it, and re-query on low confidence rather than dead-ending (`skill-discovery` rule).
78
+
68
79
  ## Capability Data Source
69
80
 
70
81
  The canonical capability matrix lives at:
@@ -1,6 +1,13 @@
1
1
  ---
2
2
  name: aiwg-writer
3
- description: Documentation persona for technical writing and content creation
3
+ description: Documentation persona/identity for technical writing, content creation, and applying voice. Use when selecting a writing persona or authoring docs in a chosen voice.
4
+ triggers:
5
+ - writer persona
6
+ - documentation persona
7
+ - technical writing voice
8
+ - select a writing identity
9
+ - soul for docs
10
+ - persona
4
11
  model: claude-sonnet-4-6
5
12
  tools:
6
13
  - Read
@@ -1,6 +1,12 @@
1
1
  ---
2
2
  name: mc-conductor
3
- description: Mission Control conductor — orchestrates parallel background missions, handles completions and failures, reports to the user
3
+ description: Mission Control conductor persona/identity — orchestrates parallel background missions, handles completions and failures, reports to the user. Use when selecting a conductor persona for mission orchestration.
4
+ triggers:
5
+ - conductor persona
6
+ - mission control identity
7
+ - orchestration conductor persona
8
+ - select a conductor persona
9
+ - persona
4
10
  model: claude-sonnet-4-6
5
11
  tools:
6
12
  - Bash
@@ -0,0 +1,112 @@
1
+ # Research Brainstorming Extension
2
+
3
+ Research brainstorming is the loose, early-stage complement to
4
+ `research-complete`. It uses wiki memory and the KB graph for ideas that are not
5
+ ready for `REF-*` source structure, GRADE assessment, or citation-grounded
6
+ synthesis.
7
+
8
+ ## Composition
9
+
10
+ This extension composes existing AIWG surfaces:
11
+
12
+ - `kb-ingest`, `aiwg kb`, and knowledge-base templates for wiki pages.
13
+ - Semantic-memory kernel skills: `memory-ingest`, `memory-query-capture`,
14
+ `memory-log-append`, and `memory-lint`.
15
+ - `aiwg index neighbors --graph kb --node <slug>` for adjacent concepts,
16
+ entities, syntheses, source summaries, and prior brainstorm notes.
17
+ - `research-query` for citation-backed corpus answers.
18
+ - `induct-research` for mature `REF-*` candidates.
19
+ - `research-store` and `.aiwg/research/working/` for promoted research tasks.
20
+
21
+ ## Corpus-Root Research Repositories
22
+
23
+ The extension also supports the proven corpus-root layout used by
24
+ `~/dev/research/research-papers`:
25
+
26
+ ```text
27
+ documentation/references/ # REF-XXX 15-section analysis files
28
+ documentation/citations/ # REF-XXX-citations sidecars and backlinks
29
+ documentation/radar/ # live freshness/GRADE radar sidecars
30
+ documentation/profiles/ # PROF-* entity profiles and graph edges
31
+ indices/ # generated cross-reference and graph indices
32
+ pdfs/full/ # archived PDFs
33
+ sources/web/ # archived web snapshots
34
+ bibliographies/master.bib # BibTeX
35
+ ```
36
+
37
+ In that layout, brainstorming stays lightweight until promotion. Promotion
38
+ targets should follow the repository contract:
39
+
40
+ - Research questions and acquisition targets stay in working notes until they
41
+ name an actual source or corpus gap.
42
+ - A promoted source candidate becomes a `REF-XXX` induction candidate and must
43
+ go through source archival, 15-section reference analysis, citation sidecar,
44
+ live radar sidecar, entity profile updates, bibliography updates, index
45
+ rebuild, and the repo's `research-lint` / `research-quality-audit` gates.
46
+ - Backlinks are mandatory: citation sidecars need incoming/outgoing consistency,
47
+ and `PROF-*` profile references need reciprocal REF links.
48
+ - Generated `indices/` are retrieval aids, not hand-authored brainstorm notes.
49
+
50
+ ## Source Labels
51
+
52
+ Every brainstorming output must label each meaningful item as one of:
53
+
54
+ - `user-idea`: supplied by the operator.
55
+ - `model-suggestion`: generated during brainstorming and not yet verified.
56
+ - `citation-backed`: grounded in an existing `REF-*`, source summary, or
57
+ research-query result.
58
+
59
+ Speculative claims stay in wiki memory until promoted. Do not require GRADE,
60
+ `REF-*`, or citation formatting for `user-idea` and `model-suggestion` notes.
61
+ Require citation and quality handling only when a note becomes a research claim,
62
+ structured synthesis, or source candidate.
63
+
64
+ ## Workflow
65
+
66
+ 1. Capture loose ideas in `.aiwg/kb/brainstorming/` using
67
+ `research-brainstorm`.
68
+ 2. Retrieve neighboring context with `research-brainstorm-query`, backed by
69
+ `aiwg index neighbors --graph kb`.
70
+ 3. Generate exploration outputs: idea maps, question trees, competing
71
+ hypotheses, synthesis sketches, and exploration plans.
72
+ 4. Promote mature material with `research-brainstorm-promote` into research
73
+ questions, acquisition targets, `REF-*` candidates, gap reports, or synthesis
74
+ report outlines.
75
+ 5. Run the normal research framework once promoted material makes a factual
76
+ research claim.
77
+
78
+ ## Current Research Grounding
79
+
80
+ The design follows current GraphRAG/KG-RAG direction: graph retrieval preserves
81
+ relationships that flat chunk retrieval loses, but incomplete graphs can mislead
82
+ retrieval and synthesis. For that reason, this extension keeps source labels,
83
+ neighbor provenance, and promotion gates explicit.
84
+
85
+ - GraphRAG survey: https://arxiv.org/abs/2501.00309
86
+ - KG-guided RAG: https://arxiv.org/abs/2502.06864
87
+ - KG-RAG under knowledge incompleteness: https://arxiv.org/abs/2504.05163
88
+ - W3C PROV overview: https://www.w3.org/TR/prov-overview/
89
+ - GRADE Working Group: https://www.gradeworkinggroup.org/
90
+
91
+ ## Storage
92
+
93
+ Default brainstorming notes live under:
94
+
95
+ ```text
96
+ .aiwg/kb/brainstorming/
97
+ ```
98
+
99
+ Promotion targets live under:
100
+
101
+ ```text
102
+ .aiwg/research/working/
103
+ .aiwg/research/reports/
104
+ .aiwg/research/findings/
105
+ ```
106
+
107
+ The KB graph remains the retrieval authority. Rebuild or query it with:
108
+
109
+ ```bash
110
+ aiwg index build --graph kb
111
+ aiwg index neighbors --graph kb --node <slug>
112
+ ```
@@ -0,0 +1,31 @@
1
+ {
2
+ "session": "local-first-command-analytics",
3
+ "seed_node": "local-first-telemetry",
4
+ "retrieval_command": "aiwg index neighbors --graph kb --node local-first-telemetry",
5
+ "neighbors": [
6
+ {
7
+ "node": "AIWG CLI",
8
+ "type": "entity",
9
+ "label": "user-idea",
10
+ "use": "Identify command surfaces that produce local events."
11
+ },
12
+ {
13
+ "node": "privacy-preserving analytics",
14
+ "type": "concept",
15
+ "label": "model-suggestion",
16
+ "use": "Frame path hashing and argument redaction questions."
17
+ },
18
+ {
19
+ "node": "command-log-design",
20
+ "type": "brainstorm-note",
21
+ "label": "user-idea",
22
+ "use": "Reuse prior notes about scope-aware storage."
23
+ }
24
+ ],
25
+ "output_shape": [
26
+ "idea map",
27
+ "question tree",
28
+ "competing hypotheses",
29
+ "exploration plan"
30
+ ]
31
+ }
@@ -0,0 +1,41 @@
1
+ # Brainstorm: Local-first command analytics
2
+
3
+ ## Metadata
4
+
5
+ - Date: 2026-06-17
6
+ - Seed: How can local command usage show which research workflows need better automation?
7
+ - Source labels allowed: user-idea, model-suggestion, citation-backed
8
+
9
+ ## User Ideas
10
+
11
+ - [user-idea] Operators repeat the same manual CLI sequences when a framework
12
+ lacks a good high-level skill.
13
+ - [user-idea] Failed commands may be more useful than successful commands for
14
+ prioritizing documentation.
15
+
16
+ ## Model Suggestions
17
+
18
+ - [model-suggestion] Group usage by project-relative path so workflows in
19
+ `.aiwg/research/` can be separated from general maintenance commands.
20
+ - [model-suggestion] Compare high-frequency command clusters against existing
21
+ skill trigger phrases.
22
+
23
+ ## Citation-Backed Context
24
+
25
+ - None yet. This note should not be treated as a research claim.
26
+
27
+ ## Open Questions
28
+
29
+ - Which commands signal research discovery versus research induction?
30
+ - What privacy constraints should apply before any heatmap feature exists?
31
+
32
+ ## Links
33
+
34
+ - Concepts: [[local-first telemetry]], [[research workflow]]
35
+ - Entities: [[AIWG CLI]]
36
+ - Prior brainstorms: []
37
+
38
+ ## Promotion Candidates
39
+
40
+ - Research question: Which local usage signals predict missing research skills?
41
+ - Acquisition target: examples of privacy-preserving local analytics in CLI tools
@@ -0,0 +1,31 @@
1
+ # Brainstorm Promotion Record
2
+
3
+ ## Source Note
4
+
5
+ - Brainstorm note: [[graph-rag-for-research-brainstorming]]
6
+ - Selection: GraphRAG may improve research brainstorming retrieval by preserving
7
+ relationships between concepts, entities, sources, and prior notes.
8
+ - Current label: citation-backed
9
+
10
+ ## Promotion Target
11
+
12
+ - Type: REF-candidate
13
+ - Path: documentation/references/REF-XXX-graph-rag-survey.md
14
+ - Layout: corpus-root
15
+
16
+ ## Rationale
17
+
18
+ The candidate names a concrete source family and is ready for the normal
19
+ research-papers induction process. Completion requires more than a note: source
20
+ archive, 15-section REF analysis, citation sidecar, live radar, profile links,
21
+ bibliography update, index rebuild, and lint/quality gates.
22
+
23
+ ## Required Follow-Up
24
+
25
+ - [ ] Archive source under `pdfs/full/` or `sources/web/`
26
+ - [ ] Create `documentation/references/REF-XXX-*.md`
27
+ - [ ] Create and backlink `documentation/citations/REF-XXX-citations.md`
28
+ - [ ] Create live `documentation/radar/REF-XXX-radar.md`
29
+ - [ ] Update `PROF-*` profile backlinks when applicable
30
+ - [ ] Update `bibliographies/master.bib`
31
+ - [ ] Rebuild `indices/` and run `research-lint` / `research-quality-audit`
@@ -0,0 +1,25 @@
1
+ # Brainstorm Promotion Record
2
+
3
+ ## Source Note
4
+
5
+ - Brainstorm note: [[local-first-command-analytics]]
6
+ - Selection: Which local usage signals predict missing research skills?
7
+ - Current label: user-idea
8
+
9
+ ## Promotion Target
10
+
11
+ - Type: research-question
12
+ - Path: .aiwg/research/working/research-questions.md
13
+ - Layout: aiwg-research
14
+
15
+ ## Rationale
16
+
17
+ The idea is specific enough to become a research question but is not yet
18
+ citation-backed. It should become an acquisition/search task before appearing in
19
+ a synthesis report.
20
+
21
+ ## Required Follow-Up
22
+
23
+ - [ ] Add citations or acquisition tasks before treating as a research claim
24
+ - [ ] Search for prior work on privacy-preserving local developer analytics
25
+ - [ ] Preserve source labels in the promoted artifact