godpowers 3.0.1 → 3.11.0

package/references/shared/DASHBOARD-CONTRACT.md

@@ -1,16 +1,52 @@
 # Godpowers Dashboard Contract
 
-This reference owns the shared dashboard shape for `/god-status`, `/god-next`, `/god-mode` closeouts, and completed command closeouts.
+This reference owns the shared closeout shape for `/god-status`, `/god-next`,
+`/god-mode` closeouts, and completed command closeouts.
 
-## Runtime source
+## Runtime Source
 
-Prefer `lib/dashboard.js` for dashboard computation. Call `dashboard.compute(projectRoot)`, then render with `dashboard.render(result)` unless the caller needs a narrower brief view.
+Prefer `lib/dashboard.js` for status computation. Call
+`dashboard.compute(projectRoot)`, then render a compact action brief unless the
+user asks for `/god-status --full` or passes a verbose flag.
 
-If the runtime module is unavailable, use a manual disk scan and say `Dashboard engine: unavailable, manual scan used`.
+If the runtime module is unavailable, use a manual disk scan quietly. Mention
+the fallback only when it changes the recommendation, then suggest
+`/god-doctor`.
 
-Never mix workflow progress with audit, hygiene, remediation, or launch-readiness scores. Label those scores separately when they appear.
+Never mix workflow progress with audit, hygiene, remediation, or
+launch-readiness scores. Label those scores separately when they appear.
 
-## Required shape
+## Default Closeout
+
+Completed work, proposals, diagnostics, lifecycle views, and status summaries
+use this shape by default:
+
+```text
+<one sentence describing the result or current position>
+
+Changed:
+- <highest-signal user-visible change>
+- <highest-signal user-visible change>
+
+Validation:
+- <command>: <result>
+
+Attention:
+- <only blockers or signals that change the recommendation>
+
+Next commands:
+- <recommended command>: <one sentence reason>
+- /god-status --full: See the complete dashboard and proactive checks.
+```
+
+Omit empty sections. Do not render rows that only say `fresh`, `clear`,
+`none`, or `not-applicable`.
+
+## Full Dashboard
+
+The full dashboard is available on request through `/god-status --full` and in
+release-gate evidence. It may also appear when a user explicitly asks for all
+status details.
 
 ```text
 Godpowers Dashboard
@@ -43,13 +79,6 @@ Deliverable progress:
   Increments: <done> done, <building> building, <pending> pending
   Ledger: .godpowers/REQUIREMENTS.md
 
-What changed:
-  1. <highest-signal user-visible change>
-  2. <highest-signal user-visible change>
-
-Validation:
-  + <command>: <result>
-
 Proactive checks:
   Checkpoint: <fresh | refreshed | missing | stale | conflicts with state.json>
   Reviews: <none | N pending, suggest /god-review-changes>
@@ -57,7 +86,7 @@ Proactive checks:
   Docs: <fresh | N stale, suggest /god-docs | repo-doc-sync ran>
   Repo surface: <fresh | N stale, suggest /god-doctor | repo-surface-sync ran>
   Host: <full | degraded | unknown with host and first gap>
-  Runtime: <not-applicable | known URL, suggest /god-test-runtime | no known URL, defer deployed verification>
+  Runtime: <not-applicable | known URL, suggest /god-test-runtime | no known URL>
   Automation: <not configured | N active | available via provider, suggest /god-automation-setup>
   Security: <clear | sensitive files changed, suggest /god-harden>
   Dependencies: <clear | dependency files changed, suggest /god-update-deps>
@@ -66,28 +95,36 @@ Proactive checks:
 Open items:
   1. <deferred staging, unstaged files, pending review, blocker, or none>
 
-Next:
-  Recommended: <one concrete command or user decision>
-  Why: <one sentence tied to current state>
+Next commands:
+- <recommended command>: <one sentence reason>
+- /god-next: Recompute the state-derived next step.
 ```
 
-## Proposition closeout
+## Next Commands
 
-When a command only recommends work, end with this proposition block unless it already launched the selected command.
+When a command only recommends work, end with this block unless it already
+launched the selected command.
 
 ```text
-Proposition:
-  1. Implement partial: <single suggested command>
-  2. Implement complete: <recipe sequence or /god-mode when safe>
-  3. Discuss more: /god-discuss <routing ambiguity or missing prerequisite>
-  4. Inspect status: /god-status, /god-locate, or /god-next
-Recommended: <one option and why>
+Next commands:
+- <best runnable command>: <plain sentence reason>
+- <second runnable command>: <plain sentence reason>
+- <third runnable command>: <plain sentence reason>
 ```
 
-## Proactive rules
+Use 1 to 4 commands. Put the recommendation first. Do not show abstract
+implementation choices when a slash command can name the next move.
+
+## Proactive Rules
 
-`/god-status` is read-only by default, so Level 3 agent work becomes a suggestion unless the user asked it to continue work.
+`/god-status` is read-only by default, so Level 3 agent work becomes a
+suggestion unless the user asked it to continue work.
 
-`/god-next` may run Level 2 local helpers only when there is a direct trigger. Standalone `/god-next` turns Level 3 work into the recommended command or a proposition option.
+`/god-next` may run Level 1 read-only checks by default. It may run Level 2
+local helpers only when there is a direct trigger and no destructive effect.
+Standalone `/god-next` turns Level 3 work into the recommended command.
 
-Report checkpoint, review, sync, docs, repo surface, host, runtime, automation, security, dependency, and hygiene signals using the labels in the required shape.
+Report checkpoint, review, sync, docs, repo surface, host, runtime, automation,
+security, dependency, and hygiene signals using the labels in the full
+dashboard only when they affect the recommendation or the user asks for full
+status.

package/references/shared/README.md

@@ -8,7 +8,7 @@ Cross-tier reference content used by multiple agents.
 - [ORCHESTRATORS.md](ORCHESTRATORS.md): composition patterns with other AI
   coding workflow systems, including migration and managed sync-back.
 - [GLOSSARY.md](GLOSSARY.md): standardized vocabulary across tiers.
-- [DASHBOARD-CONTRACT.md](DASHBOARD-CONTRACT.md): shared dashboard and proposition output contract.
+- [DASHBOARD-CONTRACT.md](DASHBOARD-CONTRACT.md): shared dashboard and `Next commands:` output contract.
 - [LOCKING.md](LOCKING.md): shared state-lock contract for mutating commands.
 
 ## Planned content

package/routing/god-demo.yaml

@@ -0,0 +1,35 @@
+apiVersion: godpowers/v1
+kind: CommandRouting
+metadata:
+  command: /god-demo
+  description: Safe sandbox proof
+  tier: 0
+  family: start
+
+prerequisites:
+  required:
+    []
+
+execution:
+  spawns: [built-in]
+  context: fresh
+  reads:
+    - fixtures/quick-proof/project/.godpowers/state.json
+    - fixtures/quick-proof/manifest.json
+  writes:
+    []
+
+success-path:
+  next-recommended: /god-first-run
+  outcome:
+    type: deterministic
+    label: Sandbox proof rendered
+    reason: The shipped quick-proof fixture can be inspected without modifying the project.
+    allowed-next: [/god-first-run, /god-init, /god-status]
+
+failure-path:
+  on-error: /god-doctor
+
+endoff:
+  state-update: tier-0 updated for /god-demo
+  events: [agent.start, agent.end]

package/routing/god-first-run.yaml

@@ -0,0 +1,34 @@
+apiVersion: godpowers/v1
+kind: CommandRouting
+metadata:
+  command: /god-first-run
+  description: Guided first 10 minute onboarding
+  tier: 0
+  family: start
+
+prerequisites:
+  required:
+    []
+
+execution:
+  spawns: [built-in]
+  context: fresh
+  reads:
+    - .godpowers/state.json
+  writes:
+    []
+
+success-path:
+  next-recommended: varies
+  outcome:
+    type: contextual
+    label: Context-specific next route
+    reason: The next route depends on whether Godpowers state exists.
+    allowed-next: [/god-demo, /god-init, /god-next, /god-help]
+
+failure-path:
+  on-error: /god-doctor
+
+endoff:
+  state-update: tier-0 updated for /god-first-run
+  events: [agent.start, agent.end]

package/routing/god-surface.yaml

@@ -0,0 +1,39 @@
+apiVersion: godpowers/v1
+kind: CommandRouting
+metadata:
+  command: /god-surface
+  description: Runtime command surface profile switcher
+  tier: 0
+  family: configure
+
+prerequisites:
+  required:
+    []
+
+execution:
+  spawns: [built-in]
+  context: fresh
+  reads:
+    - lib/install-profiles.js
+    - lib/surface-profile.js
+  writes:
+    - <runtime>/GODPOWERS_PROFILE when --apply is used
+    - <runtime>/skills/ when --apply is used
+
+standards:
+  gate-on-failure: pause-for-user
+
+success-path:
+  next-recommended: /god-help
+  outcome:
+    type: deterministic
+    label: Surface profile previewed or applied
+    reason: The selected install profile controls which slash commands are visible in the target runtime.
+    allowed-next: [/god-help, /god-help all, /god-status]
+
+failure-path:
+  on-error: /god-doctor
+
+endoff:
+  state-update: tier-0 updated for /god-surface
+  events: [agent.start, agent.end]

package/routing/recipes/try-safely.yaml

@@ -0,0 +1,26 @@
+apiVersion: godpowers/v1
+kind: Recipe
+metadata:
+  name: try-safely
+  category: starting
+  description: "Try Godpowers in a safe sandbox before touching the repo"
+
+triggers:
+  intent-keywords:
+    - "try safely"
+    - "safe demo"
+    - "show me a demo"
+    - "sandbox"
+    - "proof before setup"
+    - "first run demo"
+
+sequences:
+  default:
+    description: "Try Godpowers in a safe sandbox before touching the repo"
+    steps:
+      - command: "/god-demo"
+        why: "Runs the shipped sandbox proof without modifying the current project"
+      - command: "/god-first-run"
+        why: "Continues with guided onboarding after the proof"
+
+default-sequence: default

package/skills/god-agent-audit.md

@@ -64,12 +64,11 @@ Warnings:
 Infos: 97 missing-recommended-section suggestions across agents
        (run with --json for full list)
 
-Proposition:
-  1. Implement partial: fix the highest-priority error or warning
-  2. Implement complete: /god-agent-audit --fix for info-level placeholder sections only
-  3. Discuss more: /god-discuss agent contract findings
-  4. Inspect details: /god-agent-audit --json
-Recommended: [one option and why]
+Next commands:
+- /god-agent-audit --fix: Fix the highest-priority approved agent contract issue.
+- /god-agent-audit --fix for info-level placeholder sections only: Run the full recommended path.
+- /god-discuss agent contract findings: Resolve the open question before continuing.
+- /god-agent-audit --json: Inspect details before changing files.
 ```
 
 ## Backward-compatibility promise

package/skills/god-archaeology.md

@@ -46,10 +46,9 @@ Suggested next:
   /god-tech-debt    - assess and prioritize debt revealed
   /god-feature      - now safe to add new work with archaeology in hand
 
-Proposition:
-  1. Implement partial: /god-tech-debt for only the highest-risk areas
-  2. Implement complete: /god-reconstruct then /god-audit for full brownfield alignment
-  3. Discuss more: /god-discuss the open tribal-knowledge questions
-  4. Run God Mode: /god-mode only after reconstruction or audit makes the state clear
-Recommended: /god-reconstruct when planning artifacts are missing or stale.
+Next commands:
+- /god-tech-debt for only the highest-risk areas: Run the smallest safe next step.
+- /god-reconstruct then /god-audit for full brownfield alignment: Run the full recommended path.
+- /god-discuss the open tribal-knowledge questions: Resolve the open question before continuing.
+- /god-mode only after reconstruction or audit makes the state clear: Run the full autonomous project workflow when it fits.
 ```

package/skills/god-audit.md

@@ -52,15 +52,14 @@ After god-auditor returns:
 1. Verify AUDIT-REPORT.md exists on disk
 2. Display the summary table to the user
 3. If any artifact scored below 80%: suggest re-running the failing tier
-4. End with a proposition block:
+4. End with a Next commands block:
 
 ```
-Proposition:
-  1. Implement partial: [rerun the single lowest-scoring tier or fix the top finding]
-  2. Implement complete: [rerun all failing tiers in priority order]
-  3. Discuss more: /god-discuss audit remediation plan
-  4. Run God Mode: /god-mode only when the audit has no blocking findings
-Recommended: [one option and why]
+Next commands:
+- /god-redo [tier]: Rerun the single lowest-scoring tier or fix the top finding.
+- /god-audit: Re-run audit after failing tiers are repaired in priority order.
+- /god-discuss audit remediation plan: Resolve the open question before continuing.
+- /god-mode only when the audit has no blocking findings: Run the full autonomous project workflow when it fits.
 ```
 
 ## Output Format

package/skills/god-automation-setup.md

@@ -129,15 +129,14 @@ Approval required:
   - Confirm any host-native schedule, routine, background agent, API trigger, or connector scope
 ```
 
-## Proposition Closeout
+## Next Commands Closeout
 
 End with:
 
 ```text
-Proposition:
-  1. Implement partial: create one read-only automation after approval
-  2. Implement complete: create all safe read-only automations after approval
-  3. Discuss more: tune provider, cadence, or safety rules
-  4. Inspect status: /god-automation-status
-Recommended: <one option tied to provider capability and user risk>
+Next commands:
+- /god-automation-setup: Create one approved read-only automation.
+- /god-automation-setup --all-safe: Create all approved safe read-only automations.
+- /god-discuss automation policy: Tune provider, cadence, or safety rules.
+- /god-automation-status: Inspect status before continuing.
 ```

package/skills/god-automation-status.md

@@ -59,17 +59,16 @@ Safety rules:
   - Default templates are read-only.
 ```
 
-## Proposition Closeout
+## Next Commands Closeout
 
 End with:
 
 ```text
-Proposition:
-  1. Implement partial: /god-automation-setup for one safe read-only template
-  2. Implement complete: /god-automation-setup for all safe templates supported by the current host
-  3. Discuss more: /god-discuss automation policy and provider choice
-  4. Inspect status: /god-status to see project progress and automation status together
-Recommended: <one option tied to the detected provider>
+Next commands:
+- /god-automation-setup for one safe read-only template: Run the smallest safe next step.
+- /god-automation-setup for all safe templates supported by the current host: Run the full recommended path.
+- /god-discuss automation policy and provider choice: Resolve the open question before continuing.
+- /god-status to see project progress and automation status together: Inspect status before continuing.
 ```
 
 If no provider is available, recommend manual `/god-next` or

package/skills/god-context-scan.md

@@ -88,17 +88,16 @@ Suggested next:
 2. Accept the 2 new todos: /god-check-todos
 ```
 
-## Proposition Closeout
+## Next Commands Closeout
 
-End every context scan report with a proposition block:
+End every context scan report with a Next commands block:
 
 ```
-Proposition:
-  1. Implement partial: [first suggested action, usually /god-locate or /god-repair]
-  2. Implement complete: resolve all error and warn drifts, then rerun /god-context-scan
-  3. Discuss more: /god-discuss context drift
-  4. Inspect status: /god-status after re-orientation
-Recommended: [one option and why]
+Next commands:
+- /god-locate: Re-orient from checkpoint and disk state.
+- /god-repair: Resolve state drift before continuing.
+- /god-discuss context drift: Resolve the open question before continuing.
+- /god-status after re-orientation: Inspect status before continuing.
 ```
 
 ## What this does NOT do

package/skills/god-demo.md

@@ -0,0 +1,53 @@
+---
+name: god-demo
+description: |
+  Run a safe Godpowers sandbox proof using the shipped quick-proof fixture.
+  Shows disk-derived state, missing artifacts, host guarantees, and the next
+  command without modifying the user's project.
+
+  Triggers on: "god demo", "/god-demo", "show me a demo", "sandbox",
+  "try godpowers safely"
+---
+
+# /god-demo
+
+Show a safe sandbox proof before the user touches their project. This command
+uses the shipped quick-proof fixture and does not write to the user's project.
+
+## Process
+
+1. Resolve `<runtimeRoot>/lib/quick-proof.js`.
+2. Call `quickProof.compute(projectRoot)`.
+3. Render `quickProof.render(result, { brief: true })` by default.
+4. If the user asks for full detail, render without `brief`.
+5. End with `Next commands:`.
+
+## Output Shape
+
+```text
+The sandbox proof read a shipped fixture from disk and computed the next command without changing this repo.
+
+Changed:
+- No project files were modified.
+
+Validation:
+- quick-proof fixture: rendered disk state, missing PRD, host guarantees, and next command.
+
+Next commands:
+- /god-first-run: Continue the guided first-run path.
+- /god-init: Initialize Godpowers in this project.
+- /god-status --full: Inspect full status once a project exists.
+```
+
+## Terminal Equivalent
+
+```bash
+npx godpowers demo --project=.
+```
+
+## Rules
+
+- Never write to the current project.
+- Prefer compact proof output.
+- Keep fixture state separate from the user's project state.
+- Mention host guarantees separately from fixture progress.

package/skills/god-design-impact.md

@@ -66,12 +66,11 @@ Transitive (depth 2):
 Recommendation: Run /god-design with this change to apply.
 god-design-reviewer will gate via two-stage review.
 
-Proposition:
-  1. Implement partial: /god-design for only the lowest-risk token or component change
-  2. Implement complete: /god-design with the full proposed change
-  3. Discuss more: /god-discuss the design tradeoff before changing files
-  4. Run God Mode: /god-mode only if this design change is part of a wider project run
-Recommended: apply the smallest design change first when severity is warning
+Next commands:
+- /god-design for only the lowest-risk token or component change: Run the smallest safe next step.
+- /god-design with the full proposed change: Run the full recommended path.
+- /god-discuss the design tradeoff before changing files: Resolve the open question before continuing.
+- /god-mode only if this design change is part of a wider project run: Run the full autonomous project workflow when it fits.
 or higher.
 ```
 

package/skills/god-discuss.md

@@ -73,10 +73,9 @@ Key findings:
 
 Suggested next: [the planning command this discussion was for]
 
-Proposition:
-  1. Implement partial: [smallest safe slice or planning command]
-  2. Implement complete: [full command that consumes this brief]
-  3. Discuss more: /god-discuss [specific unresolved question]
-  4. Run God Mode: /god-mode [scope] if the user wants autonomous execution
-Recommended: [one option and why]
+Next commands:
+- /god-next: Continue with the safest command that uses this discussion.
+- /god-mode [scope]: Run the full command that consumes this brief when safe.
+- /god-discuss [specific unresolved question]: Resolve the open question before continuing.
+- /god-mode [scope] if the user wants autonomous execution: Run the full autonomous project workflow when it fits.
 ```

package/skills/god-docs.md

@@ -69,20 +69,15 @@ Godpowers may invoke docs work proactively in two ways:
 | Repo structural surface drift | Run `lib/repo-surface-sync.run` and include findings in docs scope | Do not invent routing or agent ownership prose without evidence |
 | `REVIEW-REQUIRED.md` contains docs drift items | Suggest `/god-review-changes` first | Do not auto-clear review items |
 
-When auto-invoked, show:
+When auto-invoked, show a concise default note:
 
 ```text
-Auto-invoked:
-  Trigger: docs and code changed in the same workflow
-  Agent: god-docs-writer
-  Local syncs:
-    + repo-doc-sync: <safe mechanical fixes, prose review needed, or no-op>
-    + repo-surface-sync: <structural surface fresh, scoped findings, or no-op>
-    + docs-drift-check: <N claims checked, N drift items>
-  Artifacts: .godpowers/docs/UPDATE-LOG.md or no-op
-  Log: .godpowers/docs/UPDATE-LOG.md
+Checked docs against changed code. Details were written to .godpowers/docs/UPDATE-LOG.md.
 ```
 
+Use a detailed `Auto-invoked:` card only with `--verbose` or release-gate
+debugging.
+
 ## Have-Nots
 
 Docs FAIL if:

package/skills/god-doctor.md

@@ -48,7 +48,7 @@ Plain-text report grouped by severity:
 GODPOWERS DOCTOR
 
 Install: claude (~/.claude/)
-  [OK] 117 skills installed
+  [OK] 120 skills installed
   [OK] 40 agents installed
   [OK] VERSION matches (2.3.1)
   [WARN] routing/god-doctor.yaml exists but skill file did not until now
@@ -68,17 +68,16 @@ Suggested next steps:
   2. /god-restore  (review trash)
 ```
 
-## Proposition Closeout
+## Next Commands Closeout
 
-End every human-readable doctor report with a proposition block:
+End every human-readable doctor report with a Next commands block:
 
 ```
-Proposition:
-  1. Implement partial: [safest single fix or diagnostic follow-up]
-  2. Implement complete: /god-doctor --fix when all proposed fixes are safe categories
-  3. Discuss more: /god-discuss [highest-risk warning or unclear repair]
-  4. Inspect status: /god-status after repair
-Recommended: [one option and why it is safe]
+Next commands:
+- /god-doctor --fix: Run the safest approved repair category.
+- /god-doctor --fix when all proposed fixes are safe categories: Run the full recommended path.
+- /god-discuss [highest-risk warning or unclear repair]: Resolve the open question before continuing.
+- /god-status after repair: Inspect status before continuing.
 ```
 
 If the report contains errors that need manual repair, do not recommend

package/skills/god-dogfood.md

@@ -27,12 +27,12 @@ Run deterministic messy-repo scenarios before release or after automation change
 1. Resolve the runtime root and load `lib/dogfood-runner.js`.
 2. Run `dogfood.runAll()` against `fixtures/dogfood/`.
 3. Report each scenario with pass or fail status.
-4. If any scenario fails, auto-invoke the matching specialist by visible card:
+4. If any scenario fails, recommend the matching specialist with a concise note:
    - `god-greenfieldifier` for planning-system import failures.
    - `god-context-writer` for host capability or install surface failures.
    - `god-coordinator` for extension authoring or suite release failures.
 5. Do not edit user projects while running fixture scenarios.
-6. End with the Godpowers Dashboard and make `/god-repair` the recommended
+6. End with a compact action brief and make `/god-repair` the recommended
    route when dogfood is red.
 
 ## CLI Equivalent
@@ -50,14 +50,11 @@ npx godpowers dogfood --json
 - [DECISION] The dogfood suite includes extension scaffold validation.
 - [DECISION] The dogfood suite includes a Mode D suite release dry-run.
 
-## Auto-Invoke Card
+## Automatic Work Note
 
 ```text
-Auto-invoked:
-  Trigger: /god-dogfood scenario failure
-  Agent: <god-greenfieldifier | god-context-writer | god-coordinator | none, local runtime only>
-  Local syncs:
-    + dogfood-runner: <pass, fail, or skipped reason>
-  Artifacts: fixture-only unless repair is explicitly requested
-  Log: none
+Dogfood scenarios finished. Fixture details stayed in the run output unless repair was explicitly requested.
 ```
+
+Use a detailed `Auto-invoked:` card only with `--verbose` or release-gate
+debugging.

package/skills/god-explore.md

@@ -48,11 +48,9 @@ Idea explored. Suggested framing:
 Suggested next: /god-init (commit to this framing) or /god-explore again
                 (if you want to try another angle)
 
-Proposition:
-  1. Implement partial: /god-init with this framing and stop after PRD
-  2. Implement complete: /god-mode with this framing for the full project run
-  3. Discuss more: /god-explore again with the weakest assumption
-  4. Run God Mode: /god-mode if the user is ready to build now
-Recommended: /god-init when the framing feels stable enough to turn into
-requirements.
+Next commands:
+- /god-init with this framing and stop after PRD: Run the smallest safe next step.
+- /god-mode with this framing for the full project run: Run the full recommended path.
+- /god-explore again with the weakest assumption: Resolve the open question before continuing.
+- /god-mode if the user is ready to build now: Run the full autonomous project workflow when it fits.
 ```