godpowers 3.0.1 → 3.11.0

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 (67) hide show
  1. package/CHANGELOG.md +255 -2
  2. package/README.md +31 -14
  3. package/RELEASE.md +21 -33
  4. package/SKILL.md +71 -112
  5. package/bin/install.js +44 -0
  6. package/fixtures/gate/harden-pass/.godpowers/state.json +26 -0
  7. package/lib/README.md +1 -0
  8. package/lib/artifact-map.js +2 -1
  9. package/lib/cli-dispatch.js +449 -3
  10. package/lib/command-families.js +10 -2
  11. package/lib/evidence/.provenance.json +45 -0
  12. package/lib/evidence-import.js +147 -0
  13. package/lib/evidence.js +908 -0
  14. package/lib/gate.js +26 -15
  15. package/lib/install-profiles.js +4 -1
  16. package/lib/installer-args.js +241 -1
  17. package/lib/quarterback.js +183 -0
  18. package/lib/surface-profile.js +168 -0
  19. package/lib/work-report.js +137 -0
  20. package/package.json +2 -2
  21. package/references/orchestration/GOD-MODE-RUNBOOK.md +9 -14
  22. package/references/orchestration/GOD-ORCHESTRATOR-RUNBOOK.md +40 -82
  23. package/references/shared/DASHBOARD-CONTRACT.md +66 -29
  24. package/references/shared/README.md +1 -1
  25. package/routing/god-demo.yaml +35 -0
  26. package/routing/god-first-run.yaml +34 -0
  27. package/routing/god-surface.yaml +39 -0
  28. package/routing/recipes/try-safely.yaml +26 -0
  29. package/skills/god-agent-audit.md +5 -6
  30. package/skills/god-archaeology.md +5 -6
  31. package/skills/god-audit.md +6 -7
  32. package/skills/god-automation-setup.md +6 -7
  33. package/skills/god-automation-status.md +6 -7
  34. package/skills/god-context-scan.md +7 -8
  35. package/skills/god-demo.md +53 -0
  36. package/skills/god-design-impact.md +5 -6
  37. package/skills/god-discuss.md +5 -6
  38. package/skills/god-docs.md +5 -10
  39. package/skills/god-doctor.md +8 -9
  40. package/skills/god-dogfood.md +7 -10
  41. package/skills/god-explore.md +5 -7
  42. package/skills/god-first-run.md +64 -0
  43. package/skills/god-harden.md +5 -2
  44. package/skills/god-help.md +77 -51
  45. package/skills/god-hygiene.md +5 -6
  46. package/skills/god-lifecycle.md +11 -13
  47. package/skills/god-list-assumptions.md +7 -8
  48. package/skills/god-locate.md +7 -8
  49. package/skills/god-map-codebase.md +5 -6
  50. package/skills/god-migrate.md +6 -15
  51. package/skills/god-next.md +16 -17
  52. package/skills/god-preflight.md +7 -8
  53. package/skills/god-progress.md +7 -8
  54. package/skills/god-reconcile.md +5 -6
  55. package/skills/god-reconstruct.md +5 -7
  56. package/skills/god-refactor.md +6 -7
  57. package/skills/god-roadmap-check.md +6 -7
  58. package/skills/god-scan.md +5 -9
  59. package/skills/god-spike.md +5 -6
  60. package/skills/god-standards.md +5 -6
  61. package/skills/god-status.md +12 -10
  62. package/skills/god-surface.md +61 -0
  63. package/skills/god-sync.md +4 -8
  64. package/skills/god-tech-debt.md +5 -6
  65. package/skills/god-test-runtime.md +4 -8
  66. package/skills/god-version.md +1 -1
  67. package/skills/god.md +53 -52
@@ -17,19 +17,15 @@ god-build, god-feature, god-hotfix etc. trigger automatically (per
17
17
  Phase 7 workflow integration). Useful when you want to refresh state
18
18
  between automatic triggers, or after manual code edits.
19
19
 
20
- Because `/god-scan` calls the local reverse-sync runtime directly, it must be
21
- explicit when no agent is spawned:
20
+ Because `/god-scan` calls the local reverse-sync runtime directly, keep the
21
+ default message concise:
22
22
 
23
23
  ```
24
- Auto-invoked:
25
- Trigger: /god-scan
26
- Agent: none, local runtime only
27
- Local syncs:
28
- + reverse-sync: pending
29
- Artifacts: pending
30
- Log: REVIEW-REQUIRED.md if drift is found
24
+ Scanning code links and drift. Review items will be written to REVIEW-REQUIRED.md if found.
31
25
  ```
32
26
 
27
+ Use a detailed `Auto-invoked:` card only with `--verbose` or debugging.
28
+
33
29
  ## Forms
34
30
 
35
31
  | Form | Action |
@@ -84,12 +84,11 @@ Suggested next:
84
84
  - If rejecting: archive .godpowers/spikes/<question-slug>/
85
85
  - If unclear: /god-spike with narrower question
86
86
 
87
- Proposition:
88
- 1. Implement partial: /god-feature for the smallest proven slice
89
- 2. Implement complete: /god-feature with the full recommendation
90
- 3. Discuss more: /god-discuss the remaining uncertainty
91
- 4. Run God Mode: /god-mode only if this spike unblocks the full project run
92
- Recommended: proceed only when SPIKE.md has a clear DECISION and evidence.
87
+ Next commands:
88
+ - /god-feature for the smallest proven slice: Run the smallest safe next step.
89
+ - /god-feature with the full recommendation: Run the full recommended path.
90
+ - /god-discuss the remaining uncertainty: Resolve the open question before continuing.
91
+ - /god-mode only if this spike unblocks the full project run: Run the full autonomous project workflow when it fits.
93
92
 
94
93
  REMINDER: spike code is throwaway. Do NOT merge to main.
95
94
  ```
@@ -55,12 +55,11 @@ Suggested next:
55
55
  /god-redo prd to address failures with feedback
56
56
  /god-skip prd --reason "..." to accept-as-is
57
57
 
58
- Proposition:
59
- 1. Implement partial: /god-redo [tier] with the listed failures
60
- 2. Implement complete: fix all failures, rerun /god-standards, then continue the gate
61
- 3. Discuss more: /god-discuss standards failure
62
- 4. Skip: /god-skip [tier] --reason "..." only with an explicit reason
63
- Recommended: [one option and why]
58
+ Next commands:
59
+ - /god-redo [tier] with the listed failures: Run the smallest safe next step.
60
+ - /god-standards [tier]: Re-run the standards gate after fixing all failures.
61
+ - /god-discuss standards failure: Resolve the open question before continuing.
62
+ - /god-skip [tier] --reason "..." only with an explicit reason: Skip only with an explicit reason.
64
63
  ```
65
64
 
66
65
  ## Auto-invocation
@@ -20,6 +20,7 @@ proactive checks, blockers, planning visibility, and the next action first.
20
20
  | View | Shortcut | Purpose |
21
21
  |------|----------|---------|
22
22
  | Overview | `/god-status` | Operational state, proactive checks, and blockers. |
23
+ | Full dashboard | `/god-status --full` | Complete dashboard and every proactive check. |
23
24
  | Progress ledger | `/god-progress` | Requirement and roadmap increment completion. |
24
25
  | Lifecycle phase | `/god-status --lifecycle` | Project phase and fitting workflows. |
25
26
  | Resume location | `/god-status --locate` | Orientation from checkpoint, handoff, and disk state. |
@@ -33,20 +34,21 @@ aliases for one minor release. New workflows should use `/god-status
33
34
 
34
35
  1. Check whether `.godpowers/state.json` exists. If it does not, treat `.godpowers/PROGRESS.md` only as a generated legacy fallback; if neither exists, report that no Godpowers project was found and suggest `/god-init`.
35
36
  2. Resolve the runtime root and load `<runtimeRoot>/lib/dashboard.js`.
36
- 3. Call `dashboard.compute(projectRoot)` and render with `dashboard.render(result)`.
37
+ 3. Call `dashboard.compute(projectRoot)` and render the compact action brief by default.
37
38
  4. Prefer the MCP `status` tool when it is available, and fall back to the CLI or runtime module when it is not.
38
- 5. If `--lifecycle` is present, emphasize lifecycle phase, fitting workflows, PRD visibility, roadmap visibility, and the next route.
39
- 6. If `--locate` is present, emphasize CHECKPOINT.md, HANDOFF.md, recent events, current step, and the next route.
40
- 7. Use `.godpowers/PROGRESS.md` only as generated fallback or legacy explanation when state.json is missing.
41
- 8. Scan canonical artifact paths for PRD, design, architecture, roadmap, stack, repo, build, deploy, observe, launch, harden, sync, checkpoint, and requirements evidence.
42
- 9. Compare disk state to recorded state and flag phantom resume or untracked work.
43
- 10. Offer `/god-repair` when recorded state and disk evidence conflict.
39
+ 5. If `--full` is present, render the complete dashboard with all proactive checks.
40
+ 6. If `--lifecycle` is present, emphasize lifecycle phase, fitting workflows, PRD visibility, roadmap visibility, and the next route.
41
+ 7. If `--locate` is present, emphasize CHECKPOINT.md, HANDOFF.md, recent events, current step, and the next route.
42
+ 8. Use `.godpowers/PROGRESS.md` only as generated fallback or legacy explanation when state.json is missing.
43
+ 9. Scan canonical artifact paths for PRD, design, architecture, roadmap, stack, repo, build, deploy, observe, launch, harden, sync, checkpoint, and requirements evidence.
44
+ 10. Compare disk state to recorded state and flag phantom resume or untracked work.
45
+ 11. Offer `/god-repair` when recorded state and disk evidence conflict.
44
46
 
45
47
  ## Required reference
46
48
 
47
- Read `<runtimeRoot>/references/shared/DASHBOARD-CONTRACT.md` before rendering output. The shared contract owns the dashboard shape, proactive labels, and proposition block.
49
+ Read `<runtimeRoot>/references/shared/DASHBOARD-CONTRACT.md` before rendering output. The shared contract owns the compact status shape, full dashboard shape, proactive labels, and `Next commands:` block.
48
50
 
49
- If the runtime module is unavailable, fall back to a manual scan and say `Dashboard engine: unavailable, manual scan used`.
51
+ If the runtime module is unavailable, use a manual scan quietly and suggest `/god-doctor` only when the fallback changes the recommendation.
50
52
 
51
53
  Never mix workflow progress with audit, hygiene, remediation, or launch-readiness scores. Label those scores separately when they appear.
52
54
 
@@ -54,7 +56,7 @@ Never mix workflow progress with audit, hygiene, remediation, or launch-readines
54
56
 
55
57
  `/god-status` is read-only by default. It suggests Level 3 agents instead of spawning them unless the user asked status to continue work.
56
58
 
57
- Report checkpoint, review, sync, docs, repo surface, host, runtime, automation, security, dependency, and hygiene signals using the labels in the shared dashboard contract.
59
+ Report checkpoint, review, sync, docs, repo surface, host, runtime, automation, security, dependency, and hygiene signals only when they affect the recommendation. `/god-status --full` shows every label from the shared dashboard contract.
58
60
 
59
61
  ## Mode D awareness
60
62
 
@@ -0,0 +1,61 @@
1
+ ---
2
+ name: god-surface
3
+ description: |
4
+ Preview or apply a runtime command surface profile after install. Lets users
5
+ switch between core, builder, maintainer, suite, and full without reinstalling
6
+ from scratch.
7
+
8
+ Triggers on: "god surface", "/god-surface", "switch command surface",
9
+ "change profile", "show fewer commands"
10
+ ---
11
+
12
+ # /god-surface
13
+
14
+ Preview or apply a runtime command surface profile after install. Use the
15
+ existing installer profiles: `core`, `builder`, `maintainer`, `suite`, and
16
+ `full`.
17
+
18
+ ## Process
19
+
20
+ 1. Resolve `<runtimeRoot>/lib/surface-profile.js`.
21
+ 2. Default to dry-run preview.
22
+ 3. Require an explicit runtime target such as `--codex`, `--claude`,
23
+ `--runtime=codex`, or `--all` before applying.
24
+ 4. Apply only when the user passes `--apply`.
25
+ 5. End with `Next commands:`.
26
+
27
+ ## Examples
28
+
29
+ ```bash
30
+ /god-surface --profile=core --dry-run
31
+ /god-surface --profile=builder --runtime=codex --dry-run
32
+ /god-surface --profile=maintainer --runtime=claude --apply
33
+ ```
34
+
35
+ Terminal equivalent:
36
+
37
+ ```bash
38
+ npx godpowers surface --profile=builder --codex --global --dry-run
39
+ npx godpowers surface --profile=builder --codex --global --apply
40
+ ```
41
+
42
+ ## Output Shape
43
+
44
+ ```text
45
+ The builder profile would install the product-building command surface for Codex.
46
+
47
+ Attention:
48
+ - Dry-run only. No installed skills changed.
49
+
50
+ Next commands:
51
+ - /god-surface --profile=builder --runtime=codex --apply: Apply the previewed Codex surface.
52
+ - /god-help: Open the compact help view after switching.
53
+ - /god-help all: Show the complete catalog.
54
+ ```
55
+
56
+ ## Rules
57
+
58
+ - Do not apply a profile unless the user explicitly asks with `--apply`.
59
+ - Do not change project `.godpowers/` state.
60
+ - Do not invent new profile names.
61
+ - Keep the profile source of truth in `<runtimeRoot>/lib/install-profiles.js`.
@@ -57,18 +57,14 @@ User runs `/god-sync` after manual changes. Useful for:
57
57
  - Or: re-run reconciliation against current state to detect what changed
58
58
  - Recent commits for context
59
59
 
60
- Before spawning, show a visible auto-invoke card:
60
+ Before spawning, show a concise sync note:
61
61
 
62
62
  ```
63
- Auto-invoked:
64
- Trigger: <manual /god-sync, recipe closeout, /god-mode final sync, or other source>
65
- Agent: god-updater
66
- Local syncs:
67
- - pending: feature-awareness, reverse-sync, source-sync, repo-doc-sync, repo-surface-sync, route-quality-sync, recipe-coverage-sync, release-surface-sync, pillars-sync, checkpoint-sync, context-refresh
68
- Artifacts: pending
69
- Log: .godpowers/SYNC-LOG.md
63
+ Syncing project artifacts after this workflow. Details will be written to .godpowers/SYNC-LOG.md.
70
64
  ```
71
65
 
66
+ Use the detailed helper list only with `--verbose` or release-gate debugging.
67
+
72
68
  ## Verification
73
69
 
74
70
  After god-updater returns:
@@ -54,11 +54,10 @@ Suggested next:
54
54
  - /god-upgrade for major-version migrations
55
55
  - Re-run /god-tech-debt quarterly
56
56
 
57
- Proposition:
58
- 1. Implement partial: run the command for the top P0 item only
59
- 2. Implement complete: /god-feature or /god-refactor for the full P0 bundle
60
- 3. Discuss more: /god-discuss the highest-cost or highest-risk debt item
61
- 4. Run God Mode: /god-mode only if debt work should join the full project run
62
- Recommended: start with the top P0 item, then re-run /god-tech-debt after it
57
+ Next commands:
58
+ - /god-refactor <top-p0-item>: Address the highest-priority debt item first.
59
+ - /god-feature or /god-refactor for the full P0 bundle: Run the full recommended path.
60
+ - /god-discuss the highest-cost or highest-risk debt item: Resolve the open question before continuing.
61
+ - /god-mode only if debt work should join the full project run: Run the full autonomous project workflow when it fits.
63
62
  lands.
64
63
  ```
@@ -110,18 +110,14 @@ missing. If only a local URL is evidenced, run local verification and defer
110
110
  deployed staging verification until the user provides
111
111
  `STAGING_APP_URL=<deployed staging origin>` or reaches final sign-off.
112
112
 
113
- When auto-invoked, show:
113
+ When auto-invoked, show a concise default note:
114
114
 
115
115
  ```text
116
- Auto-invoked:
117
- Trigger: frontend-visible files changed and URL was evidenced
118
- Agent: god-browser-tester
119
- Local syncs:
120
- + runtime-verification: <audit | test | full pipeline>
121
- Artifacts: .godpowers/runtime/<run-id>/
122
- Log: .godpowers/runtime/<run-id>/summary.md
116
+ Verified the evidenced runtime target. Details were written to .godpowers/runtime/<run-id>/summary.md.
123
117
  ```
124
118
 
119
+ Use a detailed `Auto-invoked:` card only with `--verbose` or debugging.
120
+
125
121
  ## Output to events.jsonl
126
122
 
127
123
  ```json
@@ -16,7 +16,7 @@ Print version and a short capability summary.
16
16
  ```
17
17
  Godpowers v2.3.1
18
18
  Install: /Users/.../.claude/ (matches package.json)
19
- Surface: 117 skills, 40 agents, 13 workflows, 42 recipes
19
+ Surface: 120 skills, 40 agents, 13 workflows, 43 recipes
20
20
  Schema: intent.v1, state.v1, events.v1, workflow.v1, routing.v1, recipe.v1
21
21
  External integrations available: impeccable, agent-browser (others lazy)
22
22
  Feature awareness: planning-system migration, source-system sync-back, context refresh, dashboard status labels, repo documentation sync, repo surface sync, quick proof, request trace, release hardening, maintenance hardening
package/skills/god.md CHANGED
@@ -2,8 +2,9 @@
2
2
  name: god
3
3
  description: |
4
4
  Front door. Take free-text intent from the user, match it to a recipe via
5
- the Godpowers runtime recipes module, and propose the matching command sequence. If no
6
- text given, fall back to state-driven suggestion (same as /god-next Mode 3).
5
+ the Godpowers runtime recipes module, and recommend one matching command
6
+ sequence. If no text is given, show one state-derived recommendation and a
7
+ few alternatives.
7
8
 
8
9
  Triggers on: "/god", "god", "/god help", "I want to ...", "how do I ..."
9
10
  (when not matched by a more specific command)
@@ -12,10 +13,13 @@ description: |
12
13
  # /god (front door)
13
14
 
14
15
  The natural-language entry point. Users describe what they want; this skill
15
- matches the intent to a recipe and suggests the right command sequence. No
16
+ matches the intent to a recipe and recommends the right command sequence. No
16
17
  agent is spawned here. This is a thin router on top of the Godpowers runtime
17
18
  recipes module.
18
19
 
20
+ If the user only types `/god`, answer with one sentence of position, one
21
+ recommended command, and 2 to 3 alternatives. Do not show a catalog.
22
+
19
23
  ## Runtime module resolution
20
24
 
21
25
  Before calling runtime modules, resolve the Godpowers runtime root:
@@ -31,13 +35,14 @@ Recipes are scenario-shaped ("I'm coming back after a week", "production is
31
35
  broken", "add a feature during the current project run") and match free-text intent. `/god` is the
32
36
  front door that turns intent into the right slash command.
33
37
 
34
- This skill complements `/god-next` rather than replacing it:
38
+ This skill is the single front door. It points to narrower views only when the
39
+ user needs them:
35
40
 
36
41
  | Skill | Best for |
37
42
  |-------|----------|
38
43
  | `/god <free text>` | "I don't know which command, but here's what I want" |
39
- | `/god-next` | "I just finished X, what's next?" or pre-flight checks |
40
- | `/god-status` | "Where are we? what's done?" |
44
+ | `/god-next` | "Continue safely from disk state" |
45
+ | `/god-status --full` | "Show every dashboard detail" |
41
46
  | `/god-init` | "Start a project here" |
42
47
  | `/god-mode` | "Run the whole project run autonomously" |
43
48
 
@@ -110,9 +115,10 @@ Treat everything after `/god` as free text. If empty, treat as state-driven.
110
115
 
111
116
  ```
112
117
  text empty?
113
- yes -> state-driven: call <runtimeRoot>/lib/recipes.js suggestForState(projectRoot)
114
- display top 3 recipes ranked by current lifecycle phase
115
- also call <runtimeRoot>/lib/router.js suggestNext(projectRoot) for structural next
118
+ yes -> state-driven: call <runtimeRoot>/lib/dashboard.js compute(projectRoot)
119
+ display one sentence of current position
120
+ recommend the structural next command first
121
+ show at most 3 alternatives
116
122
 
117
123
  no -> intent-driven: call <runtimeRoot>/lib/recipes.js matchIntent(text, projectRoot)
118
124
  call <runtimeRoot>/lib/command-families.js classifiers for capture, work size, verification, and trigger precedence hints
@@ -127,15 +133,18 @@ text empty?
127
133
  For a single high-confidence match (score >= 10):
128
134
 
129
135
  ```
130
- Best match: <recipe.metadata.name>
131
- What this does: <recipe.metadata.description>
136
+ Recommended: <command>
137
+ Why: <one sentence tied to the user's words and disk state>
132
138
 
133
139
  Sequence:
134
140
  1. <command> <why>
135
141
  2. <command> <why>
136
142
  ...
137
143
 
138
- Run this sequence? (yes / show others / cancel)
144
+ Next commands:
145
+ - <command>: Run the recommended sequence.
146
+ - /god-help <family>: See only the relevant family.
147
+ - /god-status --full: Inspect the complete dashboard first.
139
148
  ```
140
149
 
141
150
  For multiple matches (top 3):
@@ -147,38 +156,27 @@ Top matches for "<user text>":
147
156
  2. <recipe-name> (<score>) <description>
148
157
  3. <recipe-name> (<score>) <description>
149
158
 
150
- Pick one (1/2/3) or describe more specifically.
159
+ Next commands:
160
+ - <best command>: Run the strongest match.
161
+ - /god-discuss <ambiguity>: Resolve the routing ambiguity.
162
+ - /god-help <family>: See the relevant command family.
151
163
  ```
152
164
 
153
165
  For state-driven (no text):
154
166
 
155
167
  ```
156
- Where you are: <lifecycle-phase>
157
-
158
- Structural next: <command> <why>
159
-
160
- Recipes that fit your current state:
161
- - <recipe-name> <description>
162
- - <recipe-name> <description>
168
+ You are <lifecycle-phase>; the likely next move is <command>.
163
169
 
164
- Run structural next? (yes / pick recipe / cancel)
165
- ```
166
-
167
- If the answer is exploratory and proposes an approach instead of selecting or
168
- running a command, close with:
169
-
170
- ```
171
- Proposition:
172
- 1. Implement partial: <smallest safe command or slice>
173
- 2. Implement complete: <full command sequence>
174
- 3. Discuss more: /god-discuss <topic or unresolved question>
175
- 4. Run God Mode: /god-mode <scope> if the user wants the full autonomous project run
176
- Recommended: <one option and why>
170
+ Next commands:
171
+ - <command>: <why>
172
+ - /god-next: Continue with the state-derived safe step.
173
+ - /god-status --full: Inspect every dashboard check.
174
+ - /god-help all: Show the complete catalog.
177
175
  ```
178
176
 
179
177
  Do not include a God Mode option when the request is clearly a tiny fix,
180
- single-file change, or narrow question. In that case, replace it with the
181
- smallest matching command, such as `/god-fast`, `/god-spike`, or `/god-next`.
178
+ single-file change, or narrow question. In that case, recommend the smallest
179
+ matching command, such as `/god-fast`, `/god-spike`, or `/god-next`.
182
180
 
183
181
  ### Step 4: execute or hand off
184
182
 
@@ -218,7 +216,10 @@ What this does: Production is broken now
218
216
  Sequence:
219
217
  1. /god-hotfix Skip planning, debug, fix with TDD, expedited deploy, schedule postmortem
220
218
 
221
- Run this sequence? (yes / show others / cancel)
219
+ Next commands:
220
+ - /god-hotfix: Start the expedited production fix.
221
+ - /god-help operate: See production operations commands only.
222
+ - /god-status --full: Inspect the complete dashboard first.
222
223
  ```
223
224
 
224
225
  ### Example 2: ambiguous intent (multiple matches)
@@ -228,10 +229,13 @@ User: /god parallel feature during build
228
229
 
229
230
  Top matches for "parallel feature during build":
230
231
 
231
- 1. add-feature-mid-arc-pause (10) Bigger feature during the current project run; reconcile with roadmap, pause, do feature, update roadmap, resume
232
- 2. add-feature-parallel (10) Parallel feature, do not disrupt main work
232
+ 1. feature-during-current-work (10) Bigger feature during the current project run; reconcile with roadmap, pause, do feature, update roadmap, resume
233
+ 2. parallel-feature (10) Parallel feature, do not disrupt main work
233
234
 
234
- Pick one (1/2) or describe more specifically.
235
+ Next commands:
236
+ - /god-reconcile parallel feature during build: Resolve the roadmap impact first.
237
+ - /god-feature parallel feature: Start the feature as scoped work.
238
+ - /god-discuss parallel feature routing: Clarify which path fits.
235
239
  ```
236
240
 
237
241
  ### Example 3: state-driven (no text)
@@ -239,16 +243,12 @@ Pick one (1/2) or describe more specifically.
239
243
  ```
240
244
  User: /god
241
245
 
242
- Where you are: tier-1 in progress (PRD done, ARCH pending)
243
-
244
- Structural next: /god-arch
245
- Why: PRD is complete; architecture is the next gate
246
-
247
- Recipes that fit your current state:
248
- - whats-next Show the next logical step with reason
249
- - rerun-tier Redo a previous tier with new inputs
246
+ You are in planning with the PRD done and architecture pending.
250
247
 
251
- Run /god-arch? (yes / pick recipe / cancel)
248
+ Next commands:
249
+ - /god-arch: Start the next planning gate.
250
+ - /god-status --full: Inspect all dashboard details first.
251
+ - /god-help start: See start and planning commands only.
252
252
  ```
253
253
 
254
254
  ### Example 4: no match
@@ -258,11 +258,12 @@ User: /god make me a sandwich
258
258
 
259
259
  No recipe matched. Falling back to state-driven suggestion.
260
260
 
261
- Where you are: tier-3 launched
262
- Structural next: (none, project run complete)
261
+ You are launched and no recipe matched the request.
263
262
 
264
- Suggested: /god-next show all valid next-step options
265
- or: /god-status re-derive state from disk
263
+ Next commands:
264
+ - /god-next: Continue from disk state.
265
+ - /god-status --full: Re-derive full status from disk.
266
+ - /god-discuss routing ambiguity: Clarify what you want Godpowers to do.
266
267
  ```
267
268
 
268
269
  ## Why a skill, not an agent