llm-orchestrator 1.0.4 → 1.0.5
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.
- package/.claude-plugin/plugin.json +1 -1
- package/package.json +1 -1
- package/policies/capabilities.md +2 -2
- package/protocol.md +1 -1
- package/registries/agent-roles.json +370 -1
- package/schemas/agent-roles.schema.json +131 -1
- package/workflows/config.md +1 -1
- package/workflows/feature.md +1 -1
- package/workflows/refactor.md +2 -2
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "llm-orchestrator",
|
|
3
3
|
"description": "Write /task once — it plans the work, shards it across parallel subagents, gates every phase and verifies before claiming done. Claude Code, Codex, OpenCode, Kilo.",
|
|
4
|
-
"version": "1.0.
|
|
4
|
+
"version": "1.0.5",
|
|
5
5
|
"author": {
|
|
6
6
|
"name": "Bogdan-Gabriel Torcescu",
|
|
7
7
|
"url": "https://www.linkedin.com/in/bogdantorcescu/"
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "llm-orchestrator",
|
|
3
|
-
"version": "1.0.
|
|
3
|
+
"version": "1.0.5",
|
|
4
4
|
"description": "Write /task once — it plans the work, shards it across parallel subagents, gates every phase and verifies before claiming done. Claude Code, Codex, OpenCode, Kilo.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"engines": {
|
package/policies/capabilities.md
CHANGED
|
@@ -40,9 +40,9 @@ README.
|
|
|
40
40
|
| Task type | Mandatory | Optional (trigger) | Available |
|
|
41
41
|
|---|---|---|---|
|
|
42
42
|
| **INCIDENT** | `incident-start` → `-evidence` → `-fix` → `-verify` → `-close`, in that order; telemetry collection **before any hypothesis** | edge/CDN observability MCP (edge or worker symptom); native-platform debugger skill (native crash); memory recall of a prior incident (the symptom rhymes) | incident orchestration workflow, adversarial-skeptic |
|
|
43
|
-
| **FEATURE** | brainstorming before plan mode; the test-engineer writes the **failing test first** | `context7` (any library/SDK API surface — even familiar ones); design-system + UI-styling skills (any new UI); payment-provider best-practices skill (money path); semantic-code MCP such as `serena` (large cross-file edit) | code-architect role, spec-complete code generator (only when the spec is complete and no clarification is needed) |
|
|
43
|
+
| **FEATURE** | brainstorming before plan mode; the test-engineer writes the **failing test first** | `context7` (any library/SDK API surface — even familiar ones); design-system + UI-styling skills (any new UI); payment-provider best-practices skill (money path); semantic-code MCP such as `serena` (large cross-file edit) | `feature-dev-code-architect` (a harness-provided planning agent, not a core role), spec-complete code generator (only when the spec is complete and no clarification is needed) |
|
|
44
44
|
| **BUG_FIX** | systematic-debugging; **reproduce before fixing** | route-data-flow-tracer (symptom crosses layers); db-concurrency-specialist (interleaving suspected); browser automation such as `playwright` (UI-visible) | compact investigator scout, read-only exploration agent |
|
|
45
|
-
| **REFACTOR** | coverage exists **before the first edit** | code-simplifier; semantic-code MCP (symbol-level moves across many files) | bounded ≤2-file builder,
|
|
45
|
+
| **REFACTOR** | coverage exists **before the first edit** | code-simplifier; semantic-code MCP (symbol-level moves across many files) | bounded ≤2-file builder, the `explore` role |
|
|
46
46
|
| **INVESTIGATION** | adversarial challenge of the synthesis — never ship a single-source conclusion; **`exa` whenever any part of the explanation rests on external or current facts** | semantic-code MCP; memory recall before fanning out (the answer may already be filed) | read-only exploration agent, general-purpose agent |
|
|
47
47
|
| **DEPLOY** | client-compatibility check + migrations applied + smoke; telemetry collection on the soak | native-build MCP (mobile/native build); CI/build MCP (edge or worker deploy); platform-config MCP (a third-party app config was touched) | performance profiling skills |
|
|
48
48
|
| **CONFIG** | db-migration-author for **any schema change** — never hand-write a migration | provider-webhook-specialist (payment or store webhook); platform-config MCP; provider SDK upgrade skill | infrastructure API MCPs |
|
package/protocol.md
CHANGED
|
@@ -99,7 +99,7 @@ shards only. See [dispatch](policies/dispatch.md) and [routing](policies/routing
|
|
|
99
99
|
| Type | Triggers | Flow | Core agent roles |
|
|
100
100
|
|---|---|---|---|
|
|
101
101
|
| **INCIDENT** | production error, 5xx, stuck state, webhook failure, alert | Evidence → Hypothesis → Fix → Verify → Close | production-telemetry-collector ×N + route-data-flow-tracer → adversarial-skeptic → builders → code-reviewer |
|
|
102
|
-
| **FEATURE** | "build", "add", "create", "implement" | Plan → TDD → Build → Test → Review → Verify |
|
|
102
|
+
| **FEATURE** | "build", "add", "create", "implement" | Plan → TDD → Build → Test → Review → Verify | orchestrator → test-engineer → backend-fixer / frontend-fixer → code-reviewer |
|
|
103
103
|
| **BUG_FIX** | "fix", "repair", "broken" | Reproduce → Evidence → Hypothesis → Regression → Fix → Review | investigation → test-engineer → builders → code-reviewer |
|
|
104
104
|
| **REFACTOR** | "refactor", "clean up", "extract", "simplify" | Analyze → Coverage → Incremental → Review | code-simplifier → test-engineer → builders → code-reviewer |
|
|
105
105
|
| **INVESTIGATION** | "why", "analyze", "how does", "debug" | Evidence → Synthesis → Challenge → Report | evidence collectors ×N → synthesizer → adversarial-skeptic |
|
|
@@ -1 +1,370 @@
|
|
|
1
|
-
{
|
|
1
|
+
{
|
|
2
|
+
"_attribution": "llm-orchestrator · created by Bogdan-Gabriel Torcescu · https://www.linkedin.com/in/bogdantorcescu/ · keep this credit when copying or deriving",
|
|
3
|
+
"version": 1,
|
|
4
|
+
"permission_profiles": {
|
|
5
|
+
"RO": {
|
|
6
|
+
"description": "Read-only. Investigation, evidence collection, review, telemetry.",
|
|
7
|
+
"required_access": [
|
|
8
|
+
"read",
|
|
9
|
+
"grep",
|
|
10
|
+
"glob",
|
|
11
|
+
"bash_readonly"
|
|
12
|
+
],
|
|
13
|
+
"rules": [
|
|
14
|
+
"Never edits files.",
|
|
15
|
+
"Bash limited to read-only/inspection commands (no writes, no migrations, no deploys).",
|
|
16
|
+
"Reports findings back to the dispatcher; does not apply fixes."
|
|
17
|
+
]
|
|
18
|
+
},
|
|
19
|
+
"RW": {
|
|
20
|
+
"description": "Read-write within an assigned bounded shard.",
|
|
21
|
+
"required_access": [
|
|
22
|
+
"read",
|
|
23
|
+
"grep",
|
|
24
|
+
"glob",
|
|
25
|
+
"edit",
|
|
26
|
+
"write",
|
|
27
|
+
"bash"
|
|
28
|
+
],
|
|
29
|
+
"rules": [
|
|
30
|
+
"Edits are scoped to the files/directories named in its dispatch contract.",
|
|
31
|
+
"Runs project verification for its own shard before reporting done.",
|
|
32
|
+
"Does not merge, push, or deploy unless explicitly the dispatch contract's target."
|
|
33
|
+
]
|
|
34
|
+
},
|
|
35
|
+
"ORCHESTRATOR": {
|
|
36
|
+
"description": "Plans, dispatches, and integrates; owns the overall session state.",
|
|
37
|
+
"required_access": [
|
|
38
|
+
"read",
|
|
39
|
+
"grep",
|
|
40
|
+
"glob",
|
|
41
|
+
"edit",
|
|
42
|
+
"write",
|
|
43
|
+
"bash",
|
|
44
|
+
"agent_dispatch"
|
|
45
|
+
],
|
|
46
|
+
"rules": [
|
|
47
|
+
"Never bypasses a mandatory capability without declaring the gap first.",
|
|
48
|
+
"Owns fan-out sizing, shard boundaries, and integration/merge order.",
|
|
49
|
+
"Runs final verification before reporting completion to the user."
|
|
50
|
+
]
|
|
51
|
+
}
|
|
52
|
+
},
|
|
53
|
+
"roles": [
|
|
54
|
+
{
|
|
55
|
+
"id": "orchestrator",
|
|
56
|
+
"description": "Plans work, resolves capabilities, dispatches bounded shards, integrates results.",
|
|
57
|
+
"capabilities": [
|
|
58
|
+
"orchestration.bootstrap",
|
|
59
|
+
"memory.recall",
|
|
60
|
+
"memory.checkpoint",
|
|
61
|
+
"workflow.selection",
|
|
62
|
+
"delegation.bounded"
|
|
63
|
+
],
|
|
64
|
+
"skills": [
|
|
65
|
+
"using-superpowers",
|
|
66
|
+
"brainstorming",
|
|
67
|
+
"dispatching-parallel-agents",
|
|
68
|
+
"subagent-driven-development"
|
|
69
|
+
],
|
|
70
|
+
"mcps": [
|
|
71
|
+
"sequential-thinking",
|
|
72
|
+
"mempalace"
|
|
73
|
+
],
|
|
74
|
+
"best_for": "Any nontrivial task needing more than one shard or a risk-floor review seat.",
|
|
75
|
+
"permission_profile": "ORCHESTRATOR",
|
|
76
|
+
"default_tier": "S T3"
|
|
77
|
+
},
|
|
78
|
+
{
|
|
79
|
+
"id": "production-telemetry-collector",
|
|
80
|
+
"description": "Collects production logs/metrics/traces before an incident hypothesis is formed.",
|
|
81
|
+
"capabilities": [
|
|
82
|
+
"incident.telemetry_first"
|
|
83
|
+
],
|
|
84
|
+
"skills": [
|
|
85
|
+
"incident-orchestration",
|
|
86
|
+
"incident-evidence"
|
|
87
|
+
],
|
|
88
|
+
"mcps": [],
|
|
89
|
+
"best_for": "Incident evidence gathering; never forms a fix on its own.",
|
|
90
|
+
"permission_profile": "RO",
|
|
91
|
+
"default_tier": "W T0-T1"
|
|
92
|
+
},
|
|
93
|
+
{
|
|
94
|
+
"id": "route-data-flow-tracer",
|
|
95
|
+
"description": "Traces a request/data path across layers (frontend, API, DB, provider) read-only.",
|
|
96
|
+
"capabilities": [
|
|
97
|
+
"code.locate"
|
|
98
|
+
],
|
|
99
|
+
"skills": [],
|
|
100
|
+
"mcps": [
|
|
101
|
+
"serena"
|
|
102
|
+
],
|
|
103
|
+
"best_for": "Symptoms that cross architectural layers.",
|
|
104
|
+
"permission_profile": "RO",
|
|
105
|
+
"default_tier": "W T0-T1"
|
|
106
|
+
},
|
|
107
|
+
{
|
|
108
|
+
"id": "explore",
|
|
109
|
+
"description": "Read-only breadth search across a codebase: where something is defined, what calls it, which files are involved.",
|
|
110
|
+
"capabilities": [
|
|
111
|
+
"code.locate",
|
|
112
|
+
"tool.discovery"
|
|
113
|
+
],
|
|
114
|
+
"skills": [],
|
|
115
|
+
"mcps": [],
|
|
116
|
+
"best_for": "Broad read-only location work before a decision; never edits.",
|
|
117
|
+
"permission_profile": "RO",
|
|
118
|
+
"default_tier": "W T0-T1"
|
|
119
|
+
},
|
|
120
|
+
{
|
|
121
|
+
"id": "db-concurrency-specialist",
|
|
122
|
+
"description": "Reviews transactional/locking correctness and concurrency-sensitive schema/code.",
|
|
123
|
+
"capabilities": [
|
|
124
|
+
"concurrency.review",
|
|
125
|
+
"database.schema_provenance"
|
|
126
|
+
],
|
|
127
|
+
"skills": [],
|
|
128
|
+
"mcps": [
|
|
129
|
+
"db-client"
|
|
130
|
+
],
|
|
131
|
+
"best_for": "Race conditions, stale claims, lock ordering, transactional boundaries.",
|
|
132
|
+
"permission_profile": "RO",
|
|
133
|
+
"default_tier": "X T3",
|
|
134
|
+
"review_floor": "X T4"
|
|
135
|
+
},
|
|
136
|
+
{
|
|
137
|
+
"id": "frontend-specialist",
|
|
138
|
+
"description": "Implements frontend changes that touch shared state, realtime or a native bridge, respecting shipped-client compatibility.",
|
|
139
|
+
"capabilities": [
|
|
140
|
+
"ui.design_system",
|
|
141
|
+
"test.behavioral"
|
|
142
|
+
],
|
|
143
|
+
"skills": [
|
|
144
|
+
"frontend-design",
|
|
145
|
+
"design-system"
|
|
146
|
+
],
|
|
147
|
+
"mcps": [
|
|
148
|
+
"playwright"
|
|
149
|
+
],
|
|
150
|
+
"best_for": "Complex frontend state, realtime surfaces and native-bridge implementation work.",
|
|
151
|
+
"permission_profile": "RW",
|
|
152
|
+
"default_tier": "S T3"
|
|
153
|
+
},
|
|
154
|
+
{
|
|
155
|
+
"id": "provider-webhook-specialist",
|
|
156
|
+
"description": "Implements and reviews payment/webhook provider integrations (Stripe, Apple, Google).",
|
|
157
|
+
"capabilities": [
|
|
158
|
+
"billing.provider_evidence",
|
|
159
|
+
"push.provider_evidence",
|
|
160
|
+
"review.independent"
|
|
161
|
+
],
|
|
162
|
+
"skills": [
|
|
163
|
+
"stripe-best-practices"
|
|
164
|
+
],
|
|
165
|
+
"mcps": [
|
|
166
|
+
"billing-provider-api",
|
|
167
|
+
"push-provider-api"
|
|
168
|
+
],
|
|
169
|
+
"best_for": "Webhook signature/idempotency, provider state reconciliation, refund delivery.",
|
|
170
|
+
"permission_profile": "RW",
|
|
171
|
+
"default_tier": "X T2",
|
|
172
|
+
"review_floor": "X T4"
|
|
173
|
+
},
|
|
174
|
+
{
|
|
175
|
+
"id": "adversarial-skeptic",
|
|
176
|
+
"description": "Independently challenges a conclusion, diagnosis, or diff before it ships.",
|
|
177
|
+
"capabilities": [
|
|
178
|
+
"review.adversarial",
|
|
179
|
+
"review.independent"
|
|
180
|
+
],
|
|
181
|
+
"skills": [],
|
|
182
|
+
"mcps": [],
|
|
183
|
+
"best_for": "Money, auth, migration, and frozen-build-shaped review seats.",
|
|
184
|
+
"permission_profile": "RO",
|
|
185
|
+
"default_tier": "S T3",
|
|
186
|
+
"review_floor": "X T4"
|
|
187
|
+
},
|
|
188
|
+
{
|
|
189
|
+
"id": "test-engineer",
|
|
190
|
+
"description": "Writes and maintains behavioral/regression tests.",
|
|
191
|
+
"capabilities": [
|
|
192
|
+
"test.behavioral",
|
|
193
|
+
"workflow.tdd"
|
|
194
|
+
],
|
|
195
|
+
"skills": [
|
|
196
|
+
"test-driven-development"
|
|
197
|
+
],
|
|
198
|
+
"mcps": [],
|
|
199
|
+
"best_for": "Coverage gaps, regression tests for bug fixes, refactor safety nets.",
|
|
200
|
+
"permission_profile": "RW",
|
|
201
|
+
"default_tier": "S T2"
|
|
202
|
+
},
|
|
203
|
+
{
|
|
204
|
+
"id": "db-migration-author",
|
|
205
|
+
"description": "Sole authority for authoring SQL schema migrations.",
|
|
206
|
+
"capabilities": [
|
|
207
|
+
"config.migration_author",
|
|
208
|
+
"database.migration_checks",
|
|
209
|
+
"database.schema_provenance"
|
|
210
|
+
],
|
|
211
|
+
"skills": [],
|
|
212
|
+
"mcps": [
|
|
213
|
+
"db-client"
|
|
214
|
+
],
|
|
215
|
+
"best_for": "Any new migration file; never hand-write one outside this role.",
|
|
216
|
+
"permission_profile": "RW",
|
|
217
|
+
"default_tier": "S T3",
|
|
218
|
+
"review_floor": "X T4"
|
|
219
|
+
},
|
|
220
|
+
{
|
|
221
|
+
"id": "backend-fixer",
|
|
222
|
+
"description": "Implements bounded backend changes: a bug fix behind a validated hypothesis, or a feature/config change behind a failing test.",
|
|
223
|
+
"capabilities": [
|
|
224
|
+
"workflow.debug",
|
|
225
|
+
"test.behavioral"
|
|
226
|
+
],
|
|
227
|
+
"skills": [
|
|
228
|
+
"systematic-debugging",
|
|
229
|
+
"test-driven-development"
|
|
230
|
+
],
|
|
231
|
+
"mcps": [],
|
|
232
|
+
"best_for": "Backend implementation shards — bug fixes with a validated hypothesis, and the build phase of a feature or config flow.",
|
|
233
|
+
"permission_profile": "RW",
|
|
234
|
+
"default_tier": "S T2"
|
|
235
|
+
},
|
|
236
|
+
{
|
|
237
|
+
"id": "frontend-fixer",
|
|
238
|
+
"description": "Implements bounded frontend changes: a bug fix behind a validated hypothesis, or a feature change behind a failing test.",
|
|
239
|
+
"capabilities": [
|
|
240
|
+
"workflow.debug",
|
|
241
|
+
"test.behavioral",
|
|
242
|
+
"ui.design_system"
|
|
243
|
+
],
|
|
244
|
+
"skills": [
|
|
245
|
+
"systematic-debugging",
|
|
246
|
+
"test-driven-development"
|
|
247
|
+
],
|
|
248
|
+
"mcps": [
|
|
249
|
+
"playwright"
|
|
250
|
+
],
|
|
251
|
+
"best_for": "Frontend/UI implementation shards — bug fixes with a validated hypothesis, and the build phase of a feature flow.",
|
|
252
|
+
"permission_profile": "RW",
|
|
253
|
+
"default_tier": "S T2"
|
|
254
|
+
},
|
|
255
|
+
{
|
|
256
|
+
"id": "code-simplifier",
|
|
257
|
+
"description": "Simplifies and clarifies recently changed code without changing behavior.",
|
|
258
|
+
"capabilities": [
|
|
259
|
+
"refactor.preservation",
|
|
260
|
+
"code.semantic_edit"
|
|
261
|
+
],
|
|
262
|
+
"skills": [],
|
|
263
|
+
"mcps": [
|
|
264
|
+
"serena"
|
|
265
|
+
],
|
|
266
|
+
"best_for": "Post-implementation cleanup passes.",
|
|
267
|
+
"permission_profile": "RW",
|
|
268
|
+
"default_tier": "S T2"
|
|
269
|
+
},
|
|
270
|
+
{
|
|
271
|
+
"id": "code-reviewer",
|
|
272
|
+
"description": "Performs the review pass at the task's review risk floor.",
|
|
273
|
+
"capabilities": [
|
|
274
|
+
"review.independent",
|
|
275
|
+
"workflow.review"
|
|
276
|
+
],
|
|
277
|
+
"skills": [
|
|
278
|
+
"requesting-code-review",
|
|
279
|
+
"receiving-code-review"
|
|
280
|
+
],
|
|
281
|
+
"mcps": [],
|
|
282
|
+
"best_for": "The review seat on every review-gated task.",
|
|
283
|
+
"permission_profile": "RO",
|
|
284
|
+
"default_tier": "S T2",
|
|
285
|
+
"review_floor": "S T3"
|
|
286
|
+
},
|
|
287
|
+
{
|
|
288
|
+
"id": "general",
|
|
289
|
+
"description": "General-purpose bounded worker for tasks that fit no specialist role.",
|
|
290
|
+
"capabilities": [
|
|
291
|
+
"workflow.selection",
|
|
292
|
+
"delegation.bounded"
|
|
293
|
+
],
|
|
294
|
+
"skills": [],
|
|
295
|
+
"mcps": [],
|
|
296
|
+
"best_for": "Bounded work with no specialist owner; escalate rather than widen scope.",
|
|
297
|
+
"permission_profile": "RW",
|
|
298
|
+
"default_tier": "S T2"
|
|
299
|
+
}
|
|
300
|
+
],
|
|
301
|
+
"role_aliases": {
|
|
302
|
+
"planner": {
|
|
303
|
+
"roles": [
|
|
304
|
+
"orchestrator"
|
|
305
|
+
],
|
|
306
|
+
"note": "Planning is the orchestrator itself; it is never a separate dispatched agent."
|
|
307
|
+
},
|
|
308
|
+
"investigation": {
|
|
309
|
+
"roles": [
|
|
310
|
+
"route-data-flow-tracer"
|
|
311
|
+
],
|
|
312
|
+
"note": "The read-only locate/trace pass that opens a bug fix."
|
|
313
|
+
},
|
|
314
|
+
"builders": {
|
|
315
|
+
"roles": [
|
|
316
|
+
"backend-fixer",
|
|
317
|
+
"frontend-fixer"
|
|
318
|
+
],
|
|
319
|
+
"note": "Whichever side of the stack the shard owns."
|
|
320
|
+
},
|
|
321
|
+
"worker roles": {
|
|
322
|
+
"roles": [
|
|
323
|
+
"backend-fixer",
|
|
324
|
+
"frontend-fixer",
|
|
325
|
+
"code-simplifier"
|
|
326
|
+
],
|
|
327
|
+
"note": "Mechanical W-tier shards."
|
|
328
|
+
},
|
|
329
|
+
"collectors": {
|
|
330
|
+
"roles": [
|
|
331
|
+
"route-data-flow-tracer",
|
|
332
|
+
"production-telemetry-collector"
|
|
333
|
+
],
|
|
334
|
+
"note": "Evidence gathering, fanned out; research collectors use explore."
|
|
335
|
+
},
|
|
336
|
+
"synthesizer": {
|
|
337
|
+
"roles": [
|
|
338
|
+
"orchestrator"
|
|
339
|
+
],
|
|
340
|
+
"note": "Synthesis stays with the orchestrator, which holds the whole evidence set."
|
|
341
|
+
},
|
|
342
|
+
"verifier": {
|
|
343
|
+
"roles": [
|
|
344
|
+
"test-engineer"
|
|
345
|
+
],
|
|
346
|
+
"note": "Validation and smoke phases."
|
|
347
|
+
},
|
|
348
|
+
"domain specialist": {
|
|
349
|
+
"roles": [
|
|
350
|
+
"db-concurrency-specialist",
|
|
351
|
+
"provider-webhook-specialist",
|
|
352
|
+
"frontend-specialist"
|
|
353
|
+
],
|
|
354
|
+
"note": "Whichever specialist owns the money, concurrency or security surface in question."
|
|
355
|
+
},
|
|
356
|
+
"log specialist": {
|
|
357
|
+
"roles": [
|
|
358
|
+
"production-telemetry-collector"
|
|
359
|
+
],
|
|
360
|
+
"note": "Log and telemetry evidence during an incident."
|
|
361
|
+
},
|
|
362
|
+
"db/runtime specialist": {
|
|
363
|
+
"roles": [
|
|
364
|
+
"db-concurrency-specialist",
|
|
365
|
+
"route-data-flow-tracer"
|
|
366
|
+
],
|
|
367
|
+
"note": "Database and runtime-state evidence during an incident."
|
|
368
|
+
}
|
|
369
|
+
}
|
|
370
|
+
}
|
|
@@ -1 +1,131 @@
|
|
|
1
|
-
{
|
|
1
|
+
{
|
|
2
|
+
"_attribution": "llm-orchestrator · created by Bogdan-Gabriel Torcescu · https://www.linkedin.com/in/bogdantorcescu/ · keep this credit when copying or deriving",
|
|
3
|
+
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
4
|
+
"$id": "https://raw.githubusercontent.com/tbogdan/llm-orchestrator/main/schemas/agent-roles.schema.json",
|
|
5
|
+
"title": "Agent role roster and permission profiles",
|
|
6
|
+
"type": "object",
|
|
7
|
+
"required": [
|
|
8
|
+
"version",
|
|
9
|
+
"permission_profiles",
|
|
10
|
+
"roles"
|
|
11
|
+
],
|
|
12
|
+
"properties": {
|
|
13
|
+
"version": {
|
|
14
|
+
"type": "integer",
|
|
15
|
+
"minimum": 1
|
|
16
|
+
},
|
|
17
|
+
"permission_profiles": {
|
|
18
|
+
"type": "object",
|
|
19
|
+
"additionalProperties": {
|
|
20
|
+
"type": "object",
|
|
21
|
+
"required": [
|
|
22
|
+
"description",
|
|
23
|
+
"required_access",
|
|
24
|
+
"rules"
|
|
25
|
+
],
|
|
26
|
+
"properties": {
|
|
27
|
+
"description": {
|
|
28
|
+
"type": "string"
|
|
29
|
+
},
|
|
30
|
+
"required_access": {
|
|
31
|
+
"type": "array",
|
|
32
|
+
"items": {
|
|
33
|
+
"type": "string"
|
|
34
|
+
}
|
|
35
|
+
},
|
|
36
|
+
"rules": {
|
|
37
|
+
"type": "array",
|
|
38
|
+
"items": {
|
|
39
|
+
"type": "string"
|
|
40
|
+
}
|
|
41
|
+
}
|
|
42
|
+
},
|
|
43
|
+
"additionalProperties": false
|
|
44
|
+
}
|
|
45
|
+
},
|
|
46
|
+
"roles": {
|
|
47
|
+
"type": "array",
|
|
48
|
+
"items": {
|
|
49
|
+
"type": "object",
|
|
50
|
+
"required": [
|
|
51
|
+
"id",
|
|
52
|
+
"description",
|
|
53
|
+
"capabilities",
|
|
54
|
+
"skills",
|
|
55
|
+
"mcps",
|
|
56
|
+
"best_for",
|
|
57
|
+
"permission_profile",
|
|
58
|
+
"default_tier"
|
|
59
|
+
],
|
|
60
|
+
"properties": {
|
|
61
|
+
"id": {
|
|
62
|
+
"type": "string",
|
|
63
|
+
"minLength": 1
|
|
64
|
+
},
|
|
65
|
+
"description": {
|
|
66
|
+
"type": "string"
|
|
67
|
+
},
|
|
68
|
+
"capabilities": {
|
|
69
|
+
"type": "array",
|
|
70
|
+
"items": {
|
|
71
|
+
"type": "string"
|
|
72
|
+
}
|
|
73
|
+
},
|
|
74
|
+
"skills": {
|
|
75
|
+
"type": "array",
|
|
76
|
+
"items": {
|
|
77
|
+
"type": "string"
|
|
78
|
+
}
|
|
79
|
+
},
|
|
80
|
+
"mcps": {
|
|
81
|
+
"type": "array",
|
|
82
|
+
"items": {
|
|
83
|
+
"type": "string"
|
|
84
|
+
}
|
|
85
|
+
},
|
|
86
|
+
"best_for": {
|
|
87
|
+
"type": "string"
|
|
88
|
+
},
|
|
89
|
+
"permission_profile": {
|
|
90
|
+
"type": "string"
|
|
91
|
+
},
|
|
92
|
+
"default_tier": {
|
|
93
|
+
"type": "string",
|
|
94
|
+
"pattern": "^[WSXF] T[0-5](-T[0-5])?$",
|
|
95
|
+
"description": "Tier and thinking level pair resolved by registries/routing-matrix.json."
|
|
96
|
+
},
|
|
97
|
+
"review_floor": {
|
|
98
|
+
"type": "string",
|
|
99
|
+
"pattern": "^[WSXF] T[0-5](-T[0-5])?$",
|
|
100
|
+
"description": "Minimum review pair for this role domain, where policies/routing.md states one."
|
|
101
|
+
}
|
|
102
|
+
},
|
|
103
|
+
"additionalProperties": false
|
|
104
|
+
}
|
|
105
|
+
},
|
|
106
|
+
"role_aliases": {
|
|
107
|
+
"type": "object",
|
|
108
|
+
"description": "Generic function words used in workflow role tables, each resolving to one or more concrete role ids.",
|
|
109
|
+
"additionalProperties": {
|
|
110
|
+
"type": "object",
|
|
111
|
+
"required": [
|
|
112
|
+
"roles",
|
|
113
|
+
"note"
|
|
114
|
+
],
|
|
115
|
+
"additionalProperties": false,
|
|
116
|
+
"properties": {
|
|
117
|
+
"roles": {
|
|
118
|
+
"type": "array",
|
|
119
|
+
"minItems": 1,
|
|
120
|
+
"items": {
|
|
121
|
+
"type": "string"
|
|
122
|
+
}
|
|
123
|
+
},
|
|
124
|
+
"note": {
|
|
125
|
+
"type": "string"
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
}
|
|
129
|
+
}
|
|
130
|
+
}
|
|
131
|
+
}
|
package/workflows/config.md
CHANGED
|
@@ -29,7 +29,7 @@ Infrastructure API MCPs, harness config skill.
|
|
|
29
29
|
|
|
30
30
|
| Phase | Roles | Pair |
|
|
31
31
|
|---|---|---|
|
|
32
|
-
| Plan |
|
|
32
|
+
| Plan | orchestrator | S T2 |
|
|
33
33
|
| Standard config change | backend-fixer | S T2 |
|
|
34
34
|
| DB migration (migration + provenance) | db-migration-author | S T3 |
|
|
35
35
|
| Migration on money tables | db-migration-author | S T3 implementation + **X T4 review** |
|
package/workflows/feature.md
CHANGED
|
@@ -28,7 +28,7 @@ no clarification is needed).
|
|
|
28
28
|
|
|
29
29
|
| Phase | Roles | Pair |
|
|
30
30
|
|---|---|---|
|
|
31
|
-
| Plan / design |
|
|
31
|
+
| Plan / design | orchestrator (plus `feature-dev-code-architect` where the harness has it) | S T3; X T3–T4 complex; F T4 frontier |
|
|
32
32
|
| TDD | test-engineer | S T2 |
|
|
33
33
|
| Standard implementation (API CRUD, UI components) | backend-fixer, frontend-fixer | S T2 |
|
|
34
34
|
| Complex state / realtime / native bridge | frontend-specialist | S T3 |
|
package/workflows/refactor.md
CHANGED
|
@@ -19,13 +19,13 @@
|
|
|
19
19
|
|
|
20
20
|
## Available
|
|
21
21
|
|
|
22
|
-
Bounded ≤2-file builder,
|
|
22
|
+
Bounded ≤2-file builder, the `explore` role, traceability analysis where the project uses it.
|
|
23
23
|
|
|
24
24
|
## Roles and routing by phase
|
|
25
25
|
|
|
26
26
|
| Phase | Roles | Pair |
|
|
27
27
|
|---|---|---|
|
|
28
|
-
| Analysis + incremental plan | code-simplifier,
|
|
28
|
+
| Analysis + incremental plan | orchestrator, code-simplifier, explore | S T2 |
|
|
29
29
|
| Coverage | test-engineer | S T2 |
|
|
30
30
|
| Mechanical steps (moves, scoped extractions) | worker roles | W T0 |
|
|
31
31
|
| Refactor in money or security areas | domain specialist | S T3 implementation + X T4 review |
|