godpowers 0.15.8 → 0.15.10

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.
package/CHANGELOG.md CHANGED
@@ -5,6 +5,37 @@ All notable changes to Godpowers will be documented in this file.
5
5
  The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
6
6
  and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
7
7
 
8
+ ## [0.15.10] - 2026-05-11
9
+
10
+ God Mode continuity release. Makes red verification output repair work inside
11
+ the same autonomous arc instead of a terminal summary.
12
+
13
+ ### Changed
14
+ - `/god-mode` now treats failed tests, lint, typecheck, build, or check commands
15
+ as repairable work and enters an autonomous repair loop before declaring the
16
+ arc complete.
17
+ - Build completion now requires test, lint, and typecheck/check commands to be
18
+ green when those commands exist.
19
+ - `/god-mode --yolo` now auto-runs repair loops, with Critical security findings
20
+ remaining the only unconditional pause.
21
+ - Roadmap and build planning language now uses Godpowers delivery increments
22
+ instead of preserving imported methodology terminology.
23
+
24
+ ## [0.15.9] - 2026-05-11
25
+
26
+ Early design planning release. Lets UI and product-experience projects shape
27
+ DESIGN.md before architecture.
28
+
29
+ ### Changed
30
+ - `/god-init` now records UI and product-experience signals in
31
+ `.godpowers/prep/INITIAL-FINDINGS.md`.
32
+ - `/god-prd`, `/god-next`, and `/god-mode` now route to `/god-design` after
33
+ PRD and before `/god-arch` when UI or product-experience signals are found.
34
+ - `/god-design` now requires PRD, not stack, so DESIGN.md can inform
35
+ architecture, roadmap, and stack instead of arriving after them.
36
+ - Architecture, roadmap, and stack routing and agents now read DESIGN.md and
37
+ PRODUCT.md when present.
38
+
8
39
  ## [0.15.8] - 2026-05-11
9
40
 
10
41
  Init preparation release. Documents what Godpowers found before PRD, next-step
package/README.md CHANGED
@@ -2,7 +2,7 @@
2
2
 
3
3
  [![CI](https://github.com/aihxp/godpowers/actions/workflows/ci.yml/badge.svg)](https://github.com/aihxp/godpowers/actions/workflows/ci.yml)
4
4
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
5
- [![Version](https://img.shields.io/badge/version-0.15.8-blue)](CHANGELOG.md)
5
+ [![Version](https://img.shields.io/badge/version-0.15.10-blue)](CHANGELOG.md)
6
6
  [![npm](https://img.shields.io/npm/v/godpowers.svg)](https://www.npmjs.com/package/godpowers)
7
7
 
8
8
  **Ship fast. Ship right. Ship everything. Ship accountably.**
@@ -75,6 +75,12 @@ PRD complete: .godpowers/prd/PRD.md
75
75
  Suggested next: /god-arch (design the architecture)
76
76
  ```
77
77
 
78
+ For UI or product-experience projects, PRD can route to design first:
79
+
80
+ ```
81
+ Suggested next: /god-design (shape product experience)
82
+ ```
83
+
78
84
  You can also ask any time:
79
85
 
80
86
  ```
@@ -135,13 +141,18 @@ For real-world scenarios beyond greenfield:
135
141
 
136
142
  ```
137
143
  /god-mode # Standard: pauses for real questions only
138
- /god-mode --yolo # Zero pauses. Picks every default. Full send.
144
+ /god-mode --yolo # Zero pauses except Critical security. Repairs red checks before it stops.
139
145
  /god-mode --conservative # More checkpoints
140
146
  /god-mode --from=arch # Resume from a specific tier
141
147
  /god-mode --audit # Score existing artifacts. Build nothing.
142
148
  /god-mode --dry-run # Plan everything. Build nothing.
143
149
  ```
144
150
 
151
+ `/god-mode` is not complete when it merely writes planning artifacts. It keeps
152
+ going through build, verification, repair, launch, and final sync. Red tests,
153
+ typecheck, lint, build, or check output enter the repair loop instead of being
154
+ reported as the final result.
155
+
145
156
  ## Architecture
146
157
 
147
158
  ### Slash Command + Specialist Agent Pattern
@@ -164,7 +175,7 @@ Skill updates: PROGRESS.md
164
175
  | Tier | Sub-steps | Specialists |
165
176
  |------|-----------|-------------|
166
177
  | 0: Orchestration | mode detection, scale, progress | god-orchestrator |
167
- | 1: Planning | PRD, ARCH, ROADMAP, STACK | god-pm, god-architect, god-roadmapper, god-stack-selector |
178
+ | 1: Planning | PRD, optional DESIGN, ARCH, ROADMAP, STACK | god-pm, god-designer, god-architect, god-roadmapper, god-stack-selector |
168
179
  | 2: Building | repo, plan, execute, review | god-repo-scaffolder, god-planner, god-executor, god-spec-reviewer, god-quality-reviewer |
169
180
  | 3: Shipping | deploy, observe, launch, harden | god-deploy-engineer, god-observability-engineer, god-launch-strategist, god-harden-auditor |
170
181
 
@@ -23,6 +23,9 @@ Before starting:
23
23
  context.
24
24
  - Optional: `.godpowers/prep/IMPORTED-CONTEXT.md` may exist as preparation
25
25
  context.
26
+ - Optional: `.godpowers/design/DESIGN.md` and
27
+ `.godpowers/design/PRODUCT.md` may exist as early product-experience
28
+ preparation.
26
29
 
27
30
  ## Imported Preparation Context
28
31
 
@@ -34,10 +37,17 @@ If `.godpowers/prep/IMPORTED-CONTEXT.md` exists, read its technical signals
34
37
  before drafting ARCH. Use imported architecture, integration, risk, and stack
35
38
  constraints as hypothesis-level input only.
36
39
 
40
+ If `.godpowers/design/DESIGN.md` or `.godpowers/design/PRODUCT.md` exists,
41
+ read them before drafting ARCH. Use early design to identify user-facing
42
+ surfaces, routes, component boundaries, state flows, accessibility needs, and
43
+ where architecture must preserve product experience.
44
+
37
45
  Rules:
38
46
  - Do not let imported context override PRD NFRs or Godpowers state.
39
47
  - Convert useful imported signals into ADR context, tradeoffs, or open
40
48
  questions.
49
+ - Convert useful design signals into containers, ADR context, UI boundary
50
+ descriptions, or open questions.
41
51
  - If imported context conflicts with the PRD, the PRD wins and the conflict
42
52
  becomes an `[OPEN QUESTION]`.
43
53
 
@@ -4,7 +4,7 @@ description: |
4
4
  Lifecycle owner of DESIGN.md and PRODUCT.md. Detects impeccable; if
5
5
  installed, delegates to /impeccable teach (initial) or /impeccable
6
6
  document (refresh from code). If not installed, falls back to a
7
- minimal builder using PRD/ARCH/STACK.
7
+ minimal builder using prep artifacts, PRD, and any available ARCH/STACK.
8
8
 
9
9
  Spawned by: /god-design, god-orchestrator (Tier 1, conditional on UI)
10
10
  tools: Read, Write, Edit, Bash, Grep, Glob
@@ -78,10 +78,18 @@ Cascade:
78
78
  - `references/design/UX-WRITING.md` (copy)
79
79
  - `references/design/DESIGN-ANTIPATTERNS.md` (what to avoid)
80
80
 
81
- Use PRD.md (target users, register hints), ARCH.md (UI surface),
82
- STACK.md (UI framework) to generate a starter DESIGN.md from the
83
- template, applying the rules from the references above. The output
84
- will be less polished than impeccable's, but it's not toothless: the
81
+ Read `.godpowers/prep/INITIAL-FINDINGS.md` and
82
+ `.godpowers/prep/IMPORTED-CONTEXT.md` when present, then use PRD.md
83
+ (target users, flows, register hints), plus ARCH.md and STACK.md only
84
+ when they already exist. Early mode runs after PRD and before
85
+ architecture, so do not require ARCH or STACK. In early mode, produce
86
+ an experience contract that architecture and stack can honor later:
87
+ screens, flows, component needs, interaction states, product voice,
88
+ constraints, anti-references, and visual token direction.
89
+
90
+ Generate a starter DESIGN.md from the template, applying the rules
91
+ from the references above. The output will be less polished than
92
+ impeccable's, but it's not toothless: the
85
93
  references encode our design opinions across all 7 domains
86
94
  (typography, color, spatial, motion, interaction, responsive,
87
95
  ux-writing) at shallower depth than impeccable's full skill set.
@@ -152,7 +160,8 @@ After done, return to god-orchestrator with:
152
160
  - DESIGN.md path
153
161
  - PRODUCT.md path (if produced)
154
162
  - Validation summary (errors, warnings)
155
- - Suggested next: `/god-repo` (proceed to scaffolding)
163
+ - Suggested next: `/god-arch` when design ran early, or `/god-repo` when
164
+ architecture, roadmap, and stack are already complete
156
165
 
157
166
  ## Linkage hooks
158
167
 
@@ -20,6 +20,8 @@ You receive:
20
20
  - Relevant ARCH context (only what's needed for this slice)
21
21
  - Stack DECISION (tooling)
22
22
  - The slice's dependencies (what must already exist)
23
+ - Optional repair payload: failing command, error counts, focused diagnostics,
24
+ and files implicated by a previous verification run
23
25
 
24
26
  ## TDD Sequence (mandatory)
25
27
 
@@ -54,10 +56,12 @@ For every behavior in this slice:
54
56
 
55
57
  1. Run the full test suite. All tests must pass.
56
58
  2. Run the linter. All warnings resolved.
57
- 3. Stage your changes.
58
- 4. Return control to orchestrator with:
59
+ 3. Run typecheck/check command when present. All errors resolved.
60
+ 4. Stage your changes.
61
+ 5. Return control to orchestrator with:
59
62
  - Summary of what was implemented
60
63
  - Test results
64
+ - Typecheck/check results
61
65
  - Files changed
62
66
  - Ready for two-stage review
63
67
 
@@ -73,4 +77,13 @@ happen.
73
77
  - Multiple slices touched in one execution
74
78
  - Linter warnings unresolved
75
79
  - Test suite failing (any test, not just yours)
80
+ - Typecheck/check command failing
76
81
  - Stub/placeholder code in the implementation
82
+
83
+ ## Repair Mode
84
+
85
+ If invoked with a repair payload, stay narrowly focused on the failing command.
86
+ Do not reopen PRD, ARCH, roadmap, or stack unless the diagnostic proves the
87
+ artifact is stale. Fix code, config, imports, tests, generated types, or
88
+ tooling until the command passes. If the same root failure survives 3 focused
89
+ attempts, return the smallest human-only question needed to continue.
@@ -227,18 +227,23 @@ This is the third layer of decision support:
227
227
 
228
228
  ## Detection-Driven Tier 1 Routing
229
229
 
230
- After STACK done, branch on UI presence:
230
+ After PRD and before ARCH, branch on UI or product-experience presence:
231
231
 
232
232
  1. Call `lib/design-detector.isUiProject(projectRoot)` to determine
233
233
  whether DESIGN tier is required.
234
234
  2. Call `lib/design-detector.isImpeccableInstalled(projectRoot)` to
235
235
  determine whether to delegate or fall back.
236
- 3. Persist results to `state.json.project.detection-results`.
237
- 4. If `requires-design: true`: spawn `god-designer` for DESIGN tier
238
- (Tier 1 sub-step 5). god-designer delegates to impeccable's
236
+ 3. Read `.godpowers/prep/INITIAL-FINDINGS.md`,
237
+ `.godpowers/prep/IMPORTED-CONTEXT.md`, and `.godpowers/prd/PRD.md`
238
+ for UI, workflow, journey, component, brand, accessibility, and screen
239
+ signals.
240
+ 4. Persist results to `state.json.project.detection-results`.
241
+ 5. If `requires-design: true`: spawn `god-designer` for DESIGN tier before
242
+ architecture. god-designer delegates to impeccable's
239
243
  `/impeccable teach` if available, else falls back to a minimal builder.
240
- 5. If `requires-design: false`: mark `tier-1.design.status = not-required`
241
- and `tier-1.product.status = not-required`. Advance to Tier 2.
244
+ DESIGN.md and PRODUCT.md then inform architecture, roadmap, and stack.
245
+ 6. If `requires-design: false`: mark `tier-1.design.status = not-required`
246
+ and `tier-1.product.status = not-required`. Continue to architecture.
242
247
 
243
248
  ## Linkage and Reverse-Sync
244
249
 
@@ -272,17 +277,52 @@ last known hash in state.json:
272
277
 
273
278
  ## Extended Critical-Finding Gate
274
279
 
275
- The existing critical-finding gate now also fires on:
276
- - Breaking drift findings (e.g., WCAG contrast fail, ARCH container
277
- responsibility violation, STACK dep version mismatch)
278
- - Critical impeccable findings on the diff
279
- - Lint errors from `lib/artifact-linter` on any artifact
280
- - Validator errors from `lib/have-nots-validator`
280
+ The critical-finding gate fires on:
281
+ - Critical security findings from god-harden-auditor
281
282
  - god-design-reviewer BLOCK verdicts
282
-
283
- All pause both default mode AND --yolo. Lint errors cannot be
284
- auto-resolved; they're mechanical signal that something is structurally
285
- wrong.
283
+ - Breaking drift findings that would make already-written artifacts unsafe
284
+ to trust without human context
285
+ - Artifact linter or have-nots errors that still fail after repair attempts
286
+
287
+ Only Critical security findings always pause, including under --yolo.
288
+ Everything else must first enter the autonomous repair loop below. A failed
289
+ typecheck, lint, check, unit test, generated artifact lint, or have-nots pass is
290
+ not a reason to declare the arc complete. It is work.
291
+
292
+ ## Autonomous Repair Loop
293
+
294
+ Godpowers full-arc means: plan, build, verify, repair, ship, sync. Do not stop
295
+ at "artifacts generated" when the repo is still red.
296
+
297
+ When any mechanical verification fails:
298
+ - tests, typecheck, lint, formatter, build, `bun run check`, `npm run check`,
299
+ `npm test`, `cargo test`, `go test`, or equivalent
300
+ - artifact lint or have-nots validation
301
+ - generated scaffold audit with fixable failures
302
+ - launch smoke check with deterministic reproduction
303
+
304
+ Do this:
305
+ 1. Record the exact failing command, counts, and highest-signal diagnostics in
306
+ `.godpowers/build/STATE.md` or the active tier state file.
307
+ 2. Classify the failure:
308
+ - `repairable`: code, config, type, lint, test, generated artifact, missing
309
+ dependency, bad scaffold, or stale state problem.
310
+ - `human-only`: product scope contradiction, credential missing, paid vendor
311
+ decision, legal/compliance choice, or Critical security acceptance.
312
+ 3. For `repairable`, spawn the owning agent again in repair mode with only the
313
+ failing diagnostics, touched files, relevant artifact excerpts, and the
314
+ command to re-run.
315
+ 4. Re-run the failing command after each repair attempt.
316
+ 5. Repeat until green or until the same root failure survives 3 repair attempts.
317
+ 6. If repair succeeds, continue the same `/god-mode` run. Do not hand off a
318
+ "next recommended delivery increment" while required verification is red.
319
+ 7. If the same root failure survives 3 attempts, pause with a precise blocker,
320
+ attempted fixes, and the smallest human question needed to continue.
321
+
322
+ Under `--yolo`, the repair loop auto-runs. It may commit atomic repair commits
323
+ after tests pass. If a git remote exists and the user passed an explicit push
324
+ flag or the project intent says pushing is allowed, push after the green commit
325
+ and then continue the arc. Pushing is not a terminal state.
286
326
 
287
327
  ## YOLO Behavior with Design + Linkage
288
328
 
@@ -292,7 +332,8 @@ wrong.
292
332
  | Impeccable install prompt | Pause | Auto-yes; log |
293
333
  | PRODUCT.md interview | Pause | Pause anyway (load-bearing brand) |
294
334
  | Design token defaults | Pause | Auto-pick impeccable defaults; log |
295
- | Lint errors | Pause | Pause anyway |
335
+ | Repairable mechanical failures | Repair loop | Repair loop |
336
+ | Lint errors after 3 repair attempts | Pause | Pause with diagnostics |
296
337
  | Lint warnings | Continue, log | Continue, log |
297
338
  | Drift (informational) | Continue | Continue |
298
339
  | Drift (breaking) | Pause | Pause anyway |
@@ -312,8 +353,9 @@ wrong.
312
353
  5. Verify their output exists on disk
313
354
  6. Run have-nots check on the artifact
314
355
  7. If pass: update PROGRESS.md, move to next sub-step
315
- 8. If fail: spawn the agent again with the failures, OR pause for human
316
- 9. Repeat until all tiers complete
356
+ 8. If fail and repairable: enter the autonomous repair loop
357
+ 9. If fail and human-only: pause with the smallest needed question
358
+ 10. Repeat until all tiers complete and verification is green
317
359
  ```
318
360
 
319
361
  ## Specialist Agent Routing
@@ -323,9 +365,10 @@ For single-agent sub-steps:
323
365
  | Sub-step | Spawn Agent | Reads | Writes |
324
366
  |----------|-------------|-------|--------|
325
367
  | PRD | god-pm | user intent | .godpowers/prd/PRD.md |
326
- | Architecture | god-architect | PRD | .godpowers/arch/ARCH.md |
327
- | Roadmap | god-roadmapper | PRD, ARCH | .godpowers/roadmap/ROADMAP.md |
328
- | Stack | god-stack-selector | ARCH | .godpowers/stack/DECISION.md |
368
+ | Design | god-designer | prep, PRD | .godpowers/design/DESIGN.md + .godpowers/design/PRODUCT.md |
369
+ | Architecture | god-architect | PRD, optional DESIGN | .godpowers/arch/ARCH.md |
370
+ | Roadmap | god-roadmapper | PRD, ARCH, optional DESIGN | .godpowers/roadmap/ROADMAP.md |
371
+ | Stack | god-stack-selector | ARCH, optional DESIGN | .godpowers/stack/DECISION.md |
329
372
  | Repo | god-repo-scaffolder | DECISION | .godpowers/repo/AUDIT.md + repo files |
330
373
  | Deploy | god-deploy-engineer | ARCH, build | .godpowers/deploy/STATE.md |
331
374
  | Observe | god-observability-engineer | PRD, ARCH | .godpowers/observe/STATE.md |
@@ -381,7 +424,10 @@ Move to next wave only when ALL slices in current wave are committed.
381
424
  After all waves complete:
382
425
  1. Run full test suite. All must pass.
383
426
  2. Run linter. All clean.
384
- 3. Update PROGRESS.md: Build = done
427
+ 3. Run typecheck/check command when the package exposes one. All clean.
428
+ 4. If any verification fails, run the autonomous repair loop. Do not mark
429
+ Build done and do not recommend later work while verification is red.
430
+ 5. Update PROGRESS.md: Build = done
385
431
 
386
432
  CRITICAL RULES (build phase):
387
433
  - Never skip god-spec-reviewer
@@ -389,6 +435,8 @@ CRITICAL RULES (build phase):
389
435
  - Never commit without BOTH stages passing
390
436
  - Each slice gets its own atomic commit
391
437
  - Each agent gets a fresh context (defeats context rot)
438
+ - Build cannot be `done` when test, lint, typecheck, or check commands fail
439
+ - A release blocker is a repair target, not the final answer
392
440
 
393
441
  ## Post-Launch Transition (after Tier 3 completes)
394
442
 
@@ -414,8 +462,10 @@ This step runs regardless of flags:
414
462
  - /god-mode --conservative -> sync runs (with pause for confirmation)
415
463
  - /god-mode --with-hygiene -> sync runs PLUS hygiene check
416
464
 
417
- This ensures every full-arc run leaves the project in a sync'd state. The
418
- artifact coverage is consistent across all 14 categories.
465
+ After sync, re-run the final verification commands. If any are red, return to
466
+ the autonomous repair loop. This ensures every full-arc run leaves the project
467
+ green and sync'd, not merely documented. The artifact coverage is consistent
468
+ across all 14 categories.
419
469
 
420
470
  ### Steady-State Hand-off
421
471
 
@@ -499,7 +549,8 @@ that should never be auto-accepted. If god-harden-auditor returns Critical
499
549
  findings, --yolo does NOT skip. Pause for human resolution.
500
550
 
501
551
  This is the only --yolo carve-out. All other pauses are auto-resolved with
502
- the agent's documented default.
552
+ the agent's documented default, and all repairable mechanical failures are
553
+ handled by the autonomous repair loop before the arc can be called complete.
503
554
 
504
555
  ### Pause Format
505
556
 
@@ -1,8 +1,8 @@
1
1
  ---
2
2
  name: god-planner
3
3
  description: |
4
- Build planner. Reads roadmap, selects current milestone, breaks into vertical
5
- slices, identifies dependencies, groups slices into parallel waves. Each
4
+ Build planner. Reads roadmap, selects the current delivery increment, breaks
5
+ it into vertical slices, identifies dependencies, groups slices into waves. Each
6
6
  slice plan includes exact file paths, tests-first sequence, and verification
7
7
  criteria.
8
8
 
@@ -20,10 +20,11 @@ Plan the build.
20
20
 
21
21
  ## Process
22
22
 
23
- 1. Read roadmap, identify current milestone (first non-done)
23
+ 1. Read roadmap, identify the current delivery increment (first non-done
24
+ Now item)
24
25
  2. Read ARCH for technical context
25
26
  3. Read stack DECISION for tooling
26
- 4. Break milestone into **vertical slices**:
27
+ 4. Break the delivery increment into **vertical slices**:
27
28
  - Each slice delivers ONE user-visible behavior end-to-end
28
29
  - NOT "set up the database" - that's horizontal
29
30
  - YES "user can create an account" - includes DB + API + UI for that behavior
@@ -44,7 +45,7 @@ Plan the build.
44
45
  Write `.godpowers/build/PLAN.md`:
45
46
 
46
47
  ```markdown
47
- # Build Plan: Milestone [N]
48
+ # Build Plan: Delivery Increment [N]
48
49
 
49
50
  ## Wave 1 (parallel)
50
51
  ### Slice 1.1: User can create an account
@@ -1,8 +1,8 @@
1
1
  ---
2
2
  name: god-roadmapper
3
3
  description: |
4
- Sequences work into milestones with topological dependency ordering,
5
- Now/Next/Later horizons, and observable completion gates. Gated on
4
+ Sequences work into Godpowers delivery increments with topological dependency
5
+ ordering, Now/Next/Later horizons, and observable completion gates. Gated on
6
6
  Architecture.
7
7
 
8
8
  Spawned by: /god-roadmap, god-orchestrator
@@ -20,6 +20,8 @@ Optional: `.godpowers/prep/INITIAL-FINDINGS.md` may exist as preparation
20
20
  context.
21
21
  Optional: `.godpowers/prep/IMPORTED-CONTEXT.md` may exist as preparation
22
22
  context.
23
+ Optional: `.godpowers/design/DESIGN.md` and `.godpowers/design/PRODUCT.md`
24
+ may exist as early product-experience preparation.
23
25
 
24
26
  ## Imported Preparation Context
25
27
 
@@ -27,16 +29,24 @@ If `.godpowers/prep/INITIAL-FINDINGS.md` exists, read it first for repo risks,
27
29
  existing tests, docs, CI, deploy, and suggested sequencing implications.
28
30
 
29
31
  If `.godpowers/prep/IMPORTED-CONTEXT.md` exists, read its delivery signals
30
- before sequencing milestones. Use imported GSD, Superpowers, BMAD, or similar
31
- stories and plans as hypothesis-level input only.
32
+ before sequencing work. Use imported GSD, Superpowers, BMAD, or similar
33
+ stories and plans as hypothesis-level input only. Convert imported terminology
34
+ into Godpowers vocabulary. Do not preserve imported methodology terminology in
35
+ Godpowers artifacts unless the user explicitly asked for it.
36
+
37
+ If DESIGN.md or PRODUCT.md exists, read them before sequencing delivery
38
+ increments.
39
+ Use screens, flows, components, accessibility obligations, and product voice to
40
+ shape vertical slices and acceptance gates.
32
41
 
33
42
  Rules:
34
43
  - Do not add roadmap features that are not grounded in the PRD.
35
- - Use imported milestones, stories, and done-work signals to inform ordering,
44
+ - Use imported delivery, story, and done-work signals to inform ordering,
36
45
  dependency edges, and open questions.
37
46
  - If imported context conflicts with PRD or ARCH, PRD and ARCH win.
38
47
  - Mark any imported sequencing assumption as `[HYPOTHESIS]` until confirmed by
39
48
  Godpowers artifacts or the user.
49
+ - Keep design-derived delivery increments grounded in the PRD and ARCH.
40
50
 
41
51
  ## Process
42
52
 
@@ -44,10 +54,10 @@ Rules:
44
54
  2. List all features from PRD with their priority (MUST/SHOULD/COULD)
45
55
  3. Build dependency graph from ARCH (component A depends on component B)
46
56
  4. Topologically sort
47
- 5. Group features into milestones:
48
- - Each milestone has a clear, substitution-tested goal
49
- - Each milestone has an observable completion gate
50
- - Each milestone is sized: S/M/L (no day-level precision without capacity input)
57
+ 5. Group features into delivery increments:
58
+ - Each increment has a clear, substitution-tested goal
59
+ - Each increment has an observable completion gate
60
+ - Each increment is sized: S/M/L (no day-level precision without capacity input)
51
61
  6. Assign Now/Next/Later horizons:
52
62
  - **Now**: building right now, committed
53
63
  - **Next**: planned next, flexible
@@ -62,7 +72,7 @@ as the structural starting point. Write `.godpowers/roadmap/ROADMAP.md`:
62
72
  # Roadmap
63
73
 
64
74
  ## Now
65
- ### Milestone 1: [substitution-tested name]
75
+ ### Delivery Increment 1: [substitution-tested name]
66
76
  - Goal: [what users can do when this ships]
67
77
  - Gate: [observable completion criteria]
68
78
  - Size: S/M/L
@@ -70,20 +80,20 @@ as the structural starting point. Write `.godpowers/roadmap/ROADMAP.md`:
70
80
  - Features: [from PRD]
71
81
 
72
82
  ## Next
73
- [milestones]
83
+ [delivery increments]
74
84
 
75
85
  ## Later
76
- [milestones]
86
+ [themes]
77
87
  ```
78
88
 
79
89
  ## Have-Nots
80
90
 
81
91
  Roadmap FAILS if:
82
- - Milestone goal passes substitution test
92
+ - Delivery increment goal passes substitution test
83
93
  - Completion gate is not observable
84
94
  - Feature appears that is not in the PRD
85
- - All milestones the same size (no prioritization)
86
- - No dependency edges between milestones
95
+ - All increments the same size (no prioritization)
96
+ - No dependency edges between increments
87
97
  - Day-level precision without capacity input
88
98
  - Later section is empty (no long-term vision)
89
99
 
@@ -19,6 +19,8 @@ Optional: `.godpowers/prep/INITIAL-FINDINGS.md` may exist as preparation
19
19
  context.
20
20
  Optional: `.godpowers/prep/IMPORTED-CONTEXT.md` may exist as preparation
21
21
  context.
22
+ Optional: `.godpowers/design/DESIGN.md` and `.godpowers/design/PRODUCT.md`
23
+ may exist as product-experience preparation.
22
24
 
23
25
  ## Imported Preparation Context
24
26
 
@@ -29,11 +31,17 @@ If `.godpowers/prep/IMPORTED-CONTEXT.md` exists, read its technical and stack
29
31
  signals before scoring candidates. Use imported technology choices, constraints,
30
32
  and team familiarity as hypothesis-level input only.
31
33
 
34
+ If DESIGN.md or PRODUCT.md exists, read them before scoring frontend,
35
+ component, styling, accessibility, animation, and design-token tooling
36
+ candidates.
37
+
32
38
  Rules:
33
39
  - Do not let imported stack preference override ARCH NFRs or ADRs.
34
40
  - Treat imported technology choices as candidate evidence, not decisions.
35
41
  - If imported context conflicts with ARCH, ARCH wins and the conflict becomes a
36
42
  stack open question.
43
+ - If design intent conflicts with ARCH NFRs, ARCH wins and the conflict becomes
44
+ a stack or architecture open question.
37
45
 
38
46
  ## Process
39
47
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "godpowers",
3
- "version": "0.15.8",
3
+ "version": "0.15.10",
4
4
  "description": "AI-powered development system: 104 slash commands and 38 specialist agents that take a project from raw idea to hardened production. Runs inside Claude Code, Codex, Cursor, Windsurf, Gemini, and 10+ other AI coding tools.",
5
5
  "bin": {
6
6
  "godpowers": "./bin/install.js"
@@ -19,6 +19,8 @@ execution:
19
19
  context: fresh
20
20
  reads:
21
21
  - .godpowers/prd/PRD.md
22
+ - .godpowers/design/DESIGN.md
23
+ - .godpowers/design/PRODUCT.md
22
24
  - .godpowers/prep/INITIAL-FINDINGS.md
23
25
  - .godpowers/prep/IMPORTED-CONTEXT.md
24
26
  - templates/ARCH.md
@@ -10,14 +10,18 @@ prerequisites:
10
10
  - check: file:.godpowers/PROGRESS.md
11
11
  reason: "/god-design needs an initialized project"
12
12
  auto-complete: /god-init
13
- - check: file:.godpowers/stack/DECISION.md
14
- reason: "Stack must be selected before design (UI framework drives token shape)"
15
- auto-complete: /god-stack
16
- optional:
17
13
  - check: file:.godpowers/prd/PRD.md
18
- reason: "PRD informs target users and register"
14
+ reason: "Design needs PRD user, product, and register context"
15
+ auto-complete: /god-prd
16
+ optional:
17
+ - check: file:.godpowers/prep/INITIAL-FINDINGS.md
18
+ reason: "Initial findings can show UI and product-experience signals"
19
+ - check: file:.godpowers/prep/IMPORTED-CONTEXT.md
20
+ reason: "Imported GSD, Superpowers, or BMAD context can inform design hypotheses"
19
21
  - check: file:.godpowers/arch/ARCH.md
20
22
  reason: "ARCH names the UI surface that DESIGN applies to"
23
+ - check: file:.godpowers/stack/DECISION.md
24
+ reason: "STACK can refine framework-specific token shape when already selected"
21
25
 
22
26
  execution:
23
27
  spawns: [god-designer]
@@ -47,8 +51,11 @@ standards:
47
51
  - "cards in cards" # impeccable anti-pattern
48
52
 
49
53
  success-path:
50
- next-recommended: /god-repo
51
- why: Design tokens stable; repo can scaffold with token references
54
+ next-recommended: /god-arch
55
+ why: Early product experience is defined; architecture can now respect the intended surfaces and flows
56
+ alternatives:
57
+ - command: /god-repo
58
+ when: architecture-roadmap-and-stack-already-done
52
59
 
53
60
  failure-path:
54
61
  on-error: /god-doctor
@@ -17,7 +17,7 @@ execution:
17
17
  reads:
18
18
  - .godpowers/prep/INITIAL-FINDINGS.md
19
19
  - .godpowers/prep/IMPORTED-CONTEXT.md
20
- secondary-spawns: [god-pm, god-architect, god-roadmapper, god-stack-selector, god-repo-scaffolder, god-planner, god-executor, god-spec-reviewer, god-quality-reviewer, god-deploy-engineer, god-observability-engineer, god-harden-auditor, god-launch-strategist]
20
+ secondary-spawns: [god-pm, god-designer, god-architect, god-roadmapper, god-stack-selector, god-repo-scaffolder, god-planner, god-executor, god-spec-reviewer, god-quality-reviewer, god-deploy-engineer, god-observability-engineer, god-harden-auditor, god-launch-strategist]
21
21
  writes:
22
22
  []
23
23
 
@@ -37,6 +37,13 @@ standards:
37
37
 
38
38
  success-path:
39
39
  next-recommended: /god-arch
40
+ conditional-next:
41
+ - condition: ui-detected
42
+ next: /god-design
43
+ reason: UI or product-experience signals found; write DESIGN.md before architecture
44
+ - condition: no-ui-detected
45
+ next: /god-arch
46
+ reason: No UI surface detected; architecture is the next planning gate
40
47
  alternatives:
41
48
  - command: /god-mode
42
49
  when: user-wants-autonomous-from-here
@@ -17,6 +17,8 @@ execution:
17
17
  reads:
18
18
  - .godpowers/prd/PRD.md
19
19
  - .godpowers/arch/ARCH.md
20
+ - .godpowers/design/DESIGN.md
21
+ - .godpowers/design/PRODUCT.md
20
22
  - .godpowers/prep/INITIAL-FINDINGS.md
21
23
  - .godpowers/prep/IMPORTED-CONTEXT.md
22
24
  - templates/ROADMAP.md
@@ -16,6 +16,8 @@ execution:
16
16
  context: fresh
17
17
  reads:
18
18
  - .godpowers/arch/ARCH.md
19
+ - .godpowers/design/DESIGN.md
20
+ - .godpowers/design/PRODUCT.md
19
21
  - .godpowers/prep/INITIAL-FINDINGS.md
20
22
  - .godpowers/prep/IMPORTED-CONTEXT.md
21
23
  writes:
@@ -15,8 +15,11 @@ Spawn the **god-architect** agent in a fresh context via Task tool.
15
15
 
16
16
  1. Verify `.godpowers/prd/PRD.md` exists. If not, tell user to run `/god-prd` first.
17
17
  2. Spawn god-auditor briefly to verify PRD passes have-nots. If fails, report and stop.
18
- 3. Spawn god-architect with the PRD path and full context window
19
- 4. The agent writes `.godpowers/arch/ARCH.md` and ADRs to `.godpowers/arch/adr/`
18
+ 3. If `.godpowers/design/DESIGN.md` or `.godpowers/design/PRODUCT.md` exists,
19
+ include it as input. Design is optional, but when present it informs UI
20
+ containers, routes, flows, and tradeoffs.
21
+ 4. Spawn god-architect with the PRD path and full context window
22
+ 5. The agent writes `.godpowers/arch/ARCH.md` and ADRs to `.godpowers/arch/adr/`
20
23
 
21
24
  ## Verification
22
25
 
@@ -51,13 +51,19 @@ Move to next wave only when current wave is fully committed.
51
51
  After all waves:
52
52
  1. Run full test suite. All pass.
53
53
  2. Run linter. All clean.
54
- 3. Update PROGRESS.md: Build status = done
54
+ 3. Run the package's typecheck/check command when present. All pass.
55
+ 4. If any verification command fails, do not mark Build complete. Re-enter
56
+ repair mode with the owning agent, pass the exact failing diagnostics, rerun
57
+ the command, and repeat until green or until the same root failure survives
58
+ 3 repair attempts.
59
+ 5. Update PROGRESS.md: Build status = done
55
60
 
56
61
  ## Pause Conditions
57
62
 
58
63
  Pause for user ONLY if:
59
64
  - A requirement is genuinely ambiguous (two valid implementations)
60
65
  - A test reveals a gap in PRD or ARCH that needs human resolution
66
+ - The same mechanical failure remains after 3 focused repair attempts
61
67
 
62
68
  ## On Completion
63
69
 
@@ -70,8 +76,9 @@ Alternative: /god-deploy (set up deploy pipeline, parallel-safe)
70
76
  Both can run; /god-harden is the critical path to Launch.
71
77
  ```
72
78
 
73
- If more milestones remain in the roadmap, suggest re-running /god-build for
74
- the next milestone before moving to Tier 3.
79
+ If more delivery increments remain in the roadmap, continue building the next
80
+ increment before moving to Tier 3 unless the user explicitly asked to stop
81
+ after the current increment.
75
82
 
76
83
 
77
84
  ## Locking
@@ -43,6 +43,17 @@ needs to specify a mode.
43
43
  - If any are detected, summarize useful signals into
44
44
  `.godpowers/prep/IMPORTED-CONTEXT.md` as preparation context.
45
45
  Do not treat external planning-system files as source of truth.
46
+ - Detect whether early design preparation is warranted:
47
+ - UI frameworks or app models: React, Next, Vue, Nuxt, Svelte,
48
+ SvelteKit, Astro, Remix, Angular, Solid, Flutter, Electron, Tauri
49
+ - UI surfaces: `src/components/`, `app/`, `pages/`, `routes/`,
50
+ `public/`, form-heavy flows, dashboards, editor surfaces, mobile
51
+ shells, marketing pages, or other user-facing product experience
52
+ - Imported GSD, Superpowers, or BMAD context that mentions UX, screens,
53
+ journeys, components, brand, interaction states, or visual design
54
+ - Record the result in `INITIAL-FINDINGS.md` so `/god-prd`, `/god-next`,
55
+ and `/god-mode` can place `/god-design` after PRD and before
56
+ `/god-arch` when the project needs early product-experience shape.
46
57
 
47
58
  3. **Announce findings in plain English** (no jargon):
48
59
  - Empty dir + no org context: "Detected: empty directory. Starting fresh."
@@ -121,8 +132,8 @@ needs to specify a mode.
121
132
  Always create `.godpowers/prep/INITIAL-FINDINGS.md`. This is Godpowers'
122
133
  durable answer to "what did init find in this codebase?" It captures codebase
123
134
  shape, framework and tooling signals, tests, CI, docs, AI-tool files, detected
124
- methodology systems, risk signals, and the reasoning behind the suggested next
125
- command.
135
+ methodology systems, UI or product-experience signals, risk signals, and the
136
+ reasoning behind the suggested next command.
126
137
 
127
138
  If GSD, Superpowers, BMAD, or similar planning context is detected, create
128
139
  `.godpowers/prep/IMPORTED-CONTEXT.md`. This artifact is preparation context,
@@ -50,6 +50,13 @@ You are receiving a /god-mode invocation. Your job is to spawn the
50
50
  - Instruction to read `.godpowers/prep/INITIAL-FINDINGS.md` and
51
51
  `.godpowers/prep/IMPORTED-CONTEXT.md` if present before choosing the
52
52
  first planning or build step
53
+ - Instruction to run `/god-design` after `/god-prd` and before `/god-arch`
54
+ when initial findings, imported planning context, the PRD, or the
55
+ codebase show UI or product-experience signals
56
+ - Instruction that a red test, typecheck, lint, build, or check command is
57
+ not a completed arc. It must enter the autonomous repair loop and continue
58
+ the same `/god-mode` run until green, except for Critical security or a
59
+ genuine human-only decision.
53
60
 
54
61
  6. Orchestrator runs the appropriate workflow:
55
62
  - Greenfield -> full-arc
@@ -45,6 +45,11 @@ next step is safe or why a migration-aware step such as `/god-prd`,
45
45
  `/god-map-codebase`, or `/god-mode` is better. Prep artifacts are context only;
46
46
  state.json and completed Godpowers artifacts remain authoritative.
47
47
 
48
+ If PRD is complete, DESIGN is missing, and initial findings, imported context,
49
+ the PRD, or the codebase show UI or product-experience signals, suggest
50
+ `/god-design` before `/god-arch`. DESIGN.md is early preparation for visible
51
+ workflows, not a replacement for architecture.
52
+
48
53
  ## Process for Mode 1 (post-completion)
49
54
 
50
55
  ```
@@ -57,7 +62,8 @@ Skill calls: /god-next --after=/god-prd
57
62
  Read <runtimeRoot>/routing/god-prd.yaml
58
63
  |
59
64
  v
60
- Get success-path.next-recommended (e.g., "/god-arch")
65
+ Get success-path.next-recommended (e.g., "/god-arch"), then apply any
66
+ conditional-next rule such as UI-detected -> "/god-design"
61
67
  |
62
68
  v
63
69
  Display: "PRD complete: .godpowers/prd/PRD.md
package/skills/god-prd.md CHANGED
@@ -21,6 +21,9 @@ Spawn the **god-pm** agent in a fresh context via Task tool.
21
21
  5. The agent writes `.godpowers/prd/PRD.md`
22
22
  6. The agent runs have-nots checks before declaring done
23
23
  7. If god-pm pauses for a human question: relay to user using pause format
24
+ 8. If prep artifacts or the PRD show UI or product-experience signals, route
25
+ to `/god-design` next so DESIGN.md can shape architecture. Otherwise route
26
+ to `/god-arch`.
24
27
 
25
28
  ## Verification
26
29
 
@@ -47,7 +50,7 @@ After PRD is written and have-nots pass, print:
47
50
  ```
48
51
  PRD complete: .godpowers/prd/PRD.md
49
52
 
50
- Suggested next: /god-arch (design the architecture)
53
+ Suggested next: /god-design (shape product experience) if UI/product experience is detected, otherwise /god-arch (design the architecture)
51
54
  ```
52
55
 
53
56
 
@@ -1,10 +1,10 @@
1
1
  ---
2
2
  name: god-roadmap
3
3
  description: |
4
- Sequence the work into milestones. Spawns the god-roadmapper agent in a
5
- fresh context. Gated on Architecture.
4
+ Sequence the work into Godpowers delivery increments. Spawns the
5
+ god-roadmapper agent in a fresh context. Gated on Architecture.
6
6
 
7
- Triggers on: "god roadmap", "/god-roadmap", "sequence the work", "milestone plan"
7
+ Triggers on: "god roadmap", "/god-roadmap", "sequence the work", "delivery plan"
8
8
  ---
9
9
 
10
10
  # /god-roadmap
@@ -19,6 +19,7 @@
19
19
  - [HYPOTHESIS] CI or release signals found: [paths or "none detected"].
20
20
  - [HYPOTHESIS] Deployment signals found: [paths or "none detected"].
21
21
  - [HYPOTHESIS] Documentation signals found: [paths or "none detected"].
22
+ - [HYPOTHESIS] UI or product-experience signals found: [frameworks, routes, components, workflows, design docs, or "none detected"].
22
23
 
23
24
  ## Methodology And AI Tool Signals
24
25
 
@@ -31,6 +32,7 @@
31
32
  ## Planning Implications
32
33
 
33
34
  - [HYPOTHESIS] PRD should pay attention to: [product or user signal].
35
+ - [HYPOTHESIS] Design should pay attention to: [early UX, workflow, brand, component, or interaction signal].
34
36
  - [HYPOTHESIS] Architecture should pay attention to: [technical or risk signal].
35
37
  - [HYPOTHESIS] Roadmap should pay attention to: [delivery or dependency signal].
36
38
  - [HYPOTHESIS] Stack should pay attention to: [tooling or constraint signal].
@@ -1,11 +1,11 @@
1
1
  # Roadmap
2
2
 
3
- > Every milestone goal is substitution-tested. Every completion gate is
3
+ > Every delivery increment goal is substitution-tested. Every completion gate is
4
4
  > observable. No day-level precision without capacity input.
5
5
 
6
6
  ## Now (currently building, committed)
7
7
 
8
- ### Milestone 1: [Substitution-tested goal]
8
+ ### Delivery Increment 1: [Substitution-tested goal]
9
9
  - **Goal**: [What users can DO when this ships. Specific.]
10
10
  - **Completion gate**: [Observable criterion. Not "feels done".]
11
11
  - **Size**: S / M / L
@@ -14,20 +14,20 @@
14
14
  - [PRD requirement reference]
15
15
  - [PRD requirement reference]
16
16
 
17
- ### Milestone 2: [Substitution-tested goal]
17
+ ### Delivery Increment 2: [Substitution-tested goal]
18
18
  - **Goal**: ...
19
19
  - **Completion gate**: ...
20
20
  - **Size**: S / M / L
21
- - **Depends on**: Milestone 1
21
+ - **Depends on**: Delivery Increment 1
22
22
  - **Features**: ...
23
23
 
24
24
  ## Next (planned, flexible)
25
25
 
26
- ### Milestone 3: [Goal]
26
+ ### Delivery Increment 3: [Goal]
27
27
  - **Goal**: ...
28
28
  - **Completion gate**: ...
29
29
  - **Size**: ...
30
- - **Depends on**: Milestone 2
30
+ - **Depends on**: Delivery Increment 2
31
31
 
32
32
  ## Later (intent, not commitment)
33
33
 
@@ -38,10 +38,10 @@
38
38
 
39
39
  ## Have-Nots Checklist
40
40
 
41
- - [ ] Every milestone goal fails substitution test
41
+ - [ ] Every delivery increment goal fails substitution test
42
42
  - [ ] Every completion gate is observable (not subjective)
43
43
  - [ ] No feature appears that isn't in the PRD
44
- - [ ] Milestones are NOT all the same size (real prioritization)
44
+ - [ ] Delivery increments are NOT all the same size (real prioritization)
45
45
  - [ ] Dependency edges are explicit
46
46
  - [ ] No day-level precision without capacity input
47
47
  - [ ] Later section has at least one theme
@@ -58,6 +58,9 @@ jobs:
58
58
  - uses: god-quality-reviewer@^1.0.0
59
59
  on-fail: rollback
60
60
  on-pass: commit-atomic
61
+ on-fail: repair-loop
62
+ verification:
63
+ required: [test, lint, typecheck-or-check]
61
64
 
62
65
  deploy:
63
66
  tier: 3