@ludecker/aaac 1.0.0

package/templates/cursor/skills/shared/verbs/_dispatch-utils.md

@@ -0,0 +1,90 @@
+# Verb orchestrator utilities (internal)
+
+Graph paths (`agents/`, `policies/`, `skills/`, `domains/`) are relative to **`.cursor/`**.
+
+## Lifecycle (create / update / fix)
+
+**Work SSOT:** [lifecycle/lifecycle.json](../../../aaac/lifecycle/lifecycle.json) `work_phases`  
+**Gates SSOT:** [governance/gates.json](../../../aaac/governance/gates.json)  
+**Runtime:** graph `verb_runtime` — composed on Run at dispatch  
+**Run protocol:** [run/SKILL.md](../../run/SKILL.md) + [_lifecycle.md](./_lifecycle.md)
+
+Orchestrators execute phases from the Run `pending` queue. Do not skip gates unless user explicitly approves blocked run (record in Run `decisions[]`).
+
+## Load policies (every orchestrator)
+
+Read before any phase:
+
+1. [.cursor/policies/master-rules.md](../../../policies/master-rules.md)
+2. [.cursor/policies/implementation.md](../../../policies/implementation.md)
+3. [.cursor/policies/mcp-and-deploy.md](../../../policies/mcp-and-deploy.md)
+
+## Confidence gates
+
+Before `execute`, [validation](../../validation/SKILL.md) must pass:
+
+| Dimension | Min |
+|-----------|-----|
+| architecture | 0.9 |
+| requirements | 0.8 |
+| scope | 0.8 |
+
+Below threshold → set Run `status: blocked`, `awaiting_approval: true`, **STOP**
+
+## Object maturity
+
+From `object_maturity.<object>` + `maturity_rules`:
+
+| Level | Extra phases |
+|-------|----------------|
+| protected | impact_analysis, dependency_graph, rollback (mandatory) |
+| critical | impact_analysis, dependency_graph |
+| stable | impact_analysis on fix |
+| evolving | verb lifecycle only |
+
+## Load domain inventory
+
+When `$DOMAIN` slug maps to `domains/<slug>/update/inventory/SKILL.md`:
+
+1. Read inventory **first** (constraints, out-of-scope, file map)
+2. Pass inventory constraints into discovery, investigation-lite/investigation, planning, validation, execution, verification
+
+If inventory missing and command is `fix-bug` / `create-feature` / `update-module`:
+
+- Run [module-authoring](../../module-authoring/SKILL.md) discovery to bootstrap domain, **or**
+- Tell user to use generic verb command with intent
+
+## Dependency graph
+
+Before execute: [dependency-graph](../../dependency-graph/SKILL.md) resolves [dependencies.yaml](../../../aaac/dependencies.yaml) — not inventory alone.
+
+## Load object skills
+
+From [graph.yaml](../../../aaac/graph.yaml):
+
+1. Read `commands.<cmd>.object`
+2. Load every skill in `object_skills.<object>`
+3. Load extra skills from `object_skill_verbs.<object>.<verb>` when present
+
+See [_object-skills.md](./_object-skills.md) for verb-specific notes.
+
+## Investigation depth
+
+| Verb | Skill |
+|------|-------|
+| create, update | [investigation-lite](../../investigation-lite/SKILL.md) |
+| fix, fix_mode | [investigation](../../investigation/SKILL.md) → [root-cause](../../root-cause/SKILL.md) |
+
+Never skip investigation-lite on update because "change looks small."
+
+## Pre-execute gate stack (not lifecycle work)
+
+After `plan`, before `execute` — **gate** phases from [governance/gates.json](../../../aaac/governance/gates.json):
+
+1. validate
+2. impact_analysis
+3. dependency_graph
+4. fitness_functions
+5. rollback (when required by maturity or impact)
+
+On gate fail: block Run, await human approval. See [validation](../../validation/SKILL.md).

package/templates/cursor/skills/shared/verbs/_lifecycle.md

@@ -0,0 +1,87 @@
+# Verb lifecycle (SSOT)
+
+**Lifecycle = work.** **Gates = approval/checkpoints.** They are separate SSOTs, composed at runtime.
+
+| SSOT | Path | Contains |
+|------|------|----------|
+| Work phases | [lifecycle/lifecycle.json](../../lifecycle/lifecycle.json) | `verbs.*.work_phases` |
+| Gate stacks | [governance/gates.json](../../governance/gates.json) | `stacks.*` |
+| Runtime (composed) | [graph.yaml](../../graph.yaml) `verb_runtime` | work through plan + gates + execute onward |
+| Run | [run/schema.json](../../run/schema.json) | Primary execution object |
+
+Phase id → skill mapping: **[lifecycle/phases.json](../../lifecycle/phases.json)**
+
+Responsibility layers: **[layers.md](../../layers.md)**
+
+## Mental model
+
+```text
+Run
+  work:     discover → investigate* → plan
+  gates:    validate → impact → deps → fitness → rollback
+  work:     execute → verify → report
+```
+
+Everything executes within a Run. Observability (`decisions`, `log`, `checkpoints`) lives on the Run — no standalone logging.
+
+## Phase → skill map
+
+| Phase | Kind | Skill |
+|-------|------|-------|
+| `discover` | work | discovery |
+| `investigate_lite` | work | investigation-lite |
+| `investigate` | work | investigation |
+| `root_cause` | work | root-cause |
+| `plan` | work | planning |
+| `validate` | gate | validation |
+| `impact_analysis` | gate | impact-analysis |
+| `dependency_graph` | gate | dependency-graph |
+| `fitness_functions` | gate | fitness-functions |
+| `rollback` | gate | rollback |
+| `execute` | work | execution |
+| `verify` | work | testing + verification |
+| `report` | work | reporting |
+
+## Verb work phases (from lifecycle.json)
+
+| Verb | Work | Gate stack |
+|------|------|------------|
+| create | discover → investigate_lite → plan → execute → verify → report | pre_execute |
+| update | same | pre_execute |
+| fix | discover → investigate → root_cause → plan → execute → verify → report | pre_execute |
+| review | discover → plan → report | none |
+| check | discover → report | pre_execute_minimal |
+| test | discover → plan → verify → report | none |
+| release | execute → verify → report | release |
+| remove | same as update | pre_execute |
+
+## Capability routing
+
+Objects declare **capabilities** in ontology (`object_capabilities`). The registry resolves providers:
+
+```text
+object → capability → provider (skill | mcp | expert)
+```
+
+Graph `object_skills` includes **skill-type** providers only. MCP providers are recorded on Run `capabilities_resolved` and `decisions[]`. Do not hardcode provider names in orchestrators.
+
+## Human approval
+
+Gate failure or low confidence → Run:
+
+```yaml
+status: blocked
+awaiting_approval: true
+blocked_reason: "…"
+```
+
+STOP until user approves. Record approval in `decisions[]`.
+
+## Escalation
+
+If gate fails and user has not approved:
+
+```text
+STOP — awaiting approval
+Run: {run_id}
+```

package/templates/cursor/skills/shared/verbs/_object-skills.md

@@ -0,0 +1,60 @@
+# Object → skill routing (verb orchestrators)
+
+Read `object` from graph `commands.<name>.object`. Load **capabilities** from `object_capabilities.<object>`, resolve to provider skills via [capabilities/registry.json](../../../aaac/capabilities/registry.json). Graph `object_skills` is the resolved output.
+
+Extra capabilities per verb: `object_capability_verbs` in ontology (resolved to `object_skill_verbs` in graph).
+
+Layers: **code → data → product → system**.
+
+## Base map (`object_skills`)
+
+| object | Layer | Focus skill(s) |
+|--------|-------|----------------|
+| function | code | (discovery only) |
+| component | code | component, architecture |
+| module | code | domain inventory when slug present |
+| schema | data | schema |
+| model | data | model, schema |
+| migration | data | migration, schema |
+| feature | product | workflow |
+| workflow | product | workflow |
+| integration | product | integration |
+| app | system | architecture |
+| domain | system | domain |
+| architecture | system | architecture, documentation |
+
+## Verb extras (`object_skill_verbs`)
+
+| object | verb | Extra skills |
+|--------|------|--------------|
+| module | create | module-authoring |
+| module | review | architecture |
+| feature | create | module-authoring |
+
+## Domain inventory (mandatory when slug resolves)
+
+For **`fix`**, **`update`**, **`test`**, **`fix-bug`**, **`update-module`**, **`create-feature`**, **`test-module`**:
+
+1. If `domains/<slug>/update/inventory/SKILL.md` exists → read **before** discovery
+2. If missing → [module-authoring](../module-authoring/SKILL.md) bootstrap **or** ask user to pick a generic verb-object command
+
+**Do not** load `module-authoring` on **`fix-*`** unless bootstrapping a missing domain.
+
+## Resolver overrides
+
+Domain resolvers in graph override generic verb paths for `update-module`, `update-component`, `create-feature`, `fix-bug`, `test-module`, `update-architecture`.
+
+`fix-bug-by-slug` with unknown slug → `verb-fix` with `default_object: feature` per graph.
+
+## Investigation depth
+
+| Verb | Skill |
+|------|-------|
+| create, update | [investigation-lite](../investigation-lite/SKILL.md) |
+| fix, fix_mode | [investigation](../investigation/SKILL.md) → [root-cause](../root-cause/SKILL.md) |
+
+## Pre-execute gates
+
+See [_lifecycle.md](./_lifecycle.md): validation → impact-analysis → dependency-graph → fitness-functions → rollback (when required).
+
+Utilities: [_dispatch-utils.md](./_dispatch-utils.md)

package/templates/cursor/skills/shared/verbs/check/orchestrator/SKILL.md

@@ -0,0 +1,22 @@
+---
+name: verb-check-orchestrator
+description: Orchestrates all check-* commands. Internal only.
+disable-model-invocation: true
+---
+
+# check-* orchestrator
+
+**Readonly.** **Object** from graph. Domain optional for `check-function`.
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) first.
+
+## Phases
+
+1. **policies** — load `.cursor/policies/`
+2. **scope** — restate question; narrow via domain slug, inventory, intent
+3. **load_inventory** — when domain slug maps to inventory
+4. **object_skills** — from graph `object_skills.<object>`
+5. [check](../../check/SKILL.md) — swarm per check skill
+6. [reporting](../../reporting/SKILL.md) — **Answer** (yes/no/partial) then **How**
+
+No code changes. For test runs use `test-*`; for fixes use `fix-*`.

package/templates/cursor/skills/shared/verbs/check/orchestrator/contract.yaml

@@ -0,0 +1,24 @@
+name: verb-check-orchestrator
+purpose: Validate or inspect capability per check-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded
+  - check swarm completed (3+ readonly agents)
+  - answer is yes, no, or partial with conditions
+  - evidence includes path:line citations
+  - no file edits
+failure_conditions:
+  - code changes during check
+dependencies:
+  skills: [check, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/architecture.md
+verification: []

package/templates/cursor/skills/shared/verbs/create/orchestrator/SKILL.md

@@ -0,0 +1,39 @@
+---
+name: verb-create-orchestrator
+description: Orchestrates all create-* commands. Internal only.
+disable-model-invocation: true
+---
+
+# create-* orchestrator
+
+**Object** from graph `commands.<cmd>.object`.
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) and [_lifecycle.md](../_lifecycle.md) first.
+
+Lifecycle: graph `verb_runtime.create` (work + gates on Run)
+
+## Parse
+
+- **Domain:** slug (required except `create-function` — domain optional)
+- **Intent:** quoted — what to add
+
+## Phases (deterministic — do not skip)
+
+1. **policies** — load `.cursor/policies/`
+2. **load_inventory** — when `domains/<slug>/update/inventory/SKILL.md` exists
+3. **discover** — [discovery](../../discovery/SKILL.md) readonly
+4. **investigate_lite** — [investigation-lite](../../investigation-lite/SKILL.md)
+5. **plan** — [planning](../../planning/SKILL.md)
+6. **validate** — [validation](../../validation/SKILL.md) — confidence gates
+7. **impact_analysis** — [impact-analysis](../../impact-analysis/SKILL.md)
+8. **dependency_graph** — [dependency-graph](../../dependency-graph/SKILL.md)
+9. **fitness_functions** — [fitness-functions](../../fitness-functions/SKILL.md)
+10. **rollback** — [rollback](../../rollback/SKILL.md) when object maturity protected or impact high
+11. **execute** — [execution](../../execution/SKILL.md) + object_skills + module-authoring when listed
+12. **verify** — [testing](../../testing/SKILL.md) + [verification](../../verification/SKILL.md)
+13. **sync_inventory** — when domain inventory exists and execute ran
+14. **report** — [reporting](../../reporting/SKILL.md)
+
+Domain slugs with graph resolver use domain orchestrator instead (e.g. `create-feature` → `create-feature-by-slug`).
+
+Gate failure → **STOP, REQUEST CLARIFICATION**

package/templates/cursor/skills/shared/verbs/create/orchestrator/contract.yaml

@@ -0,0 +1,34 @@
+name: verb-create-orchestrator
+purpose: Create new code, config, or docs per create-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  code_changes:
+    type: boolean
+  inventory_synced:
+    type: boolean
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded before execute
+  - domain inventory loaded when domains/<slug>/update/inventory exists
+  - object skills and module-authoring loaded when graph requires
+  - governance/implementation followed for all edits
+  - fallow check_changed clean on touched files
+failure_conditions:
+  - execute without approved plan
+  - skip module-authoring when object_skill_verbs lists it
+dependencies:
+  skills: [discovery, planning, execution, testing, verification, reporting, module-authoring]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/master_rules.md
+    - docs/architecture.md
+    - docs/design_system.md
+verification:
+  - sync_inventory
+  - run_tests
+  - fallow_check_changed

package/templates/cursor/skills/shared/verbs/fix/orchestrator/SKILL.md

@@ -0,0 +1,33 @@
+---
+name: verb-fix-orchestrator
+description: Orchestrates fix-* except fix-bug resolver paths. Internal only.
+disable-model-invocation: true
+---
+
+# fix-* orchestrator
+
+**Object** from graph. `fix-bug` uses domain resolver (see [dispatch.md](../../../aaac/dispatch.md) for fallback).
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) and [_lifecycle.md](../_lifecycle.md) first.
+
+Lifecycle: graph `verb_runtime.fix` (work + gates on Run)
+
+## Phases (deterministic — do not skip)
+
+1. **policies**
+2. **load_inventory** — when domain inventory exists
+3. **discover** — [discovery](../../discovery/SKILL.md) readonly
+4. **investigate** — [investigation](../../investigation/SKILL.md) deep
+5. **root_cause** — [root-cause](../../root-cause/SKILL.md)
+6. **plan** — [planning](../../planning/SKILL.md)
+7. **validate** — [validation](../../validation/SKILL.md)
+8. **impact_analysis** — [impact-analysis](../../impact-analysis/SKILL.md)
+9. **dependency_graph** — [dependency-graph](../../dependency-graph/SKILL.md)
+10. **fitness_functions** — [fitness-functions](../../fitness-functions/SKILL.md)
+11. **rollback** — [rollback](../../rollback/SKILL.md) when maturity protected/critical or blast_radius ≥ medium
+12. **execute** — [execution](../../execution/SKILL.md)
+13. **verify** — [testing](../../testing/SKILL.md) + [verification](../../verification/SKILL.md)
+14. **sync_inventory**
+15. **report** — [reporting](../../reporting/SKILL.md)
+
+Gate failure → **STOP, REQUEST CLARIFICATION**

package/templates/cursor/skills/shared/verbs/fix/orchestrator/contract.yaml

@@ -0,0 +1,38 @@
+name: verb-fix-orchestrator
+purpose: Repair broken code, config, or docs per fix-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  code_changes:
+    type: boolean
+  inventory_synced:
+    type: boolean
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded before execute
+  - investigation completed before planning
+  - domain inventory loaded when domains/<slug>/update/inventory exists
+  - object skills from graph object_skills / object_skill_verbs loaded
+  - governance/implementation followed for all edits
+  - fallow check_changed clean on touched files
+  - user intent satisfied per verification
+failure_conditions:
+  - execute without approved plan
+  - skip investigation phase
+  - skip governance/implementation on code changes
+  - skip domain inventory when slug resolves to inventory file
+dependencies:
+  skills: [investigation, discovery, planning, execution, testing, verification, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/master_rules.md
+    - docs/architecture.md
+    - docs/design_system.md
+verification:
+  - sync_inventory
+  - run_tests
+  - fallow_check_changed

package/templates/cursor/skills/shared/verbs/release/orchestrator/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: verb-release-orchestrator
+description: Orchestrates release-* commands. Internal only.
+disable-model-invocation: true
+---
+
+# release-* orchestrator
+
+**Object** from graph. `release-app` → [platform-release](../../platform-release/orchestrator/SKILL.md) (full platform ship).
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) first.
+
+## By object
+
+| object | Behavior |
+|--------|----------|
+| app | platform-release orchestrator (Render, git, observability) |
+| feature | Flag/experiment rollout checklist; PostHog when configured |
+| integration | Integration smoke + env verification per [integration](../../integration/SKILL.md) |
+
+## Phases (non-app)
+
+1. **policies** — load `.cursor/policies/` including mcp-and-deploy
+2. **object_skills** — from graph `object_skills.<object>`
+3. [testing](../../testing/SKILL.md) + [verification](../../verification/SKILL.md)
+4. [execution](../../execution/SKILL.md) only for deploy/config steps intent requests
+5. [reporting](../../reporting/SKILL.md)
+
+Valid release objects: `app`, `feature`, `integration`. Others rejected — see graph `invalid_pairs`.

package/templates/cursor/skills/shared/verbs/release/orchestrator/contract.yaml

@@ -0,0 +1,25 @@
+name: verb-release-orchestrator
+purpose: Deploy or ship per release-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded including mcp-and-deploy
+  - tests pass before deploy steps
+  - render.yaml blueprint respected (no manual infra creation)
+  - health checks verified for API releases
+failure_conditions:
+  - deploy without verification
+  - manual Render service creation outside blueprint
+dependencies:
+  skills: [testing, verification, execution, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/architecture.md
+verification:
+  - run_tests

package/templates/cursor/skills/shared/verbs/remove/orchestrator/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: verb-remove-orchestrator
+description: Orchestrates all remove-* commands. Internal only.
+disable-model-invocation: true
+---
+
+# remove-* orchestrator
+
+**Object** from graph.
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) first.
+
+## Phases
+
+1. **policies** — load `.cursor/policies/`
+2. **load_inventory** — when domain slug maps to inventory
+3. [discovery](../../discovery/SKILL.md) — blast radius, dependents
+4. **object_skills** — from graph `object_skills.<object>`
+5. [remove](../../remove/SKILL.md) — confirm scope in report before edits
+6. [execution](../../execution/SKILL.md) — deletions and doc/inventory updates
+7. [testing](../../testing/SKILL.md) + [verification](../../verification/SKILL.md)
+8. **sync_inventory** — when domain inventory exists
+9. [reporting](../../reporting/SKILL.md)
+
+`remove-app`: require intent to name explicit retirement scope; never delete auth/deploy SSOT without user confirmation.

package/templates/cursor/skills/shared/verbs/remove/orchestrator/contract.yaml

@@ -0,0 +1,32 @@
+name: verb-remove-orchestrator
+purpose: Delete or retire code, config, or docs per remove-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  code_changes:
+    type: boolean
+  inventory_synced:
+    type: boolean
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded before execute
+  - blast radius documented before deletions
+  - governance/implementation followed
+  - fallow check_changed clean on touched files
+failure_conditions:
+  - delete without scope confirmation in report
+  - remove auth/deploy SSOT without explicit user confirm
+dependencies:
+  skills: [discovery, remove, execution, testing, verification, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/master_rules.md
+    - docs/architecture.md
+verification:
+  - sync_inventory
+  - run_tests
+  - fallow_check_changed

package/templates/cursor/skills/shared/verbs/review/orchestrator/SKILL.md

@@ -0,0 +1,24 @@
+---
+name: verb-review-orchestrator
+description: Orchestrates review-* except review-module and review-incident. Internal only.
+disable-model-invocation: true
+---
+
+# review-* orchestrator
+
+**Readonly.** **Object** from graph.
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) first.
+
+`review-module` → architecture orchestrator. `review-incident` → investigation orchestrator.
+
+## Phases
+
+1. **policies** — load `.cursor/policies/`
+2. **load_inventory** — when domain slug maps to inventory
+3. [discovery](../../discovery/SKILL.md) — focused scope
+4. **object_skills** — from graph `object_skills.<object>` + `object_skill_verbs.<object>.review`
+5. [architecture](../../architecture/SKILL.md) when object is `module`, `component`, `app`, or `architecture`
+6. [reporting](../../reporting/SKILL.md) — findings prioritized critical vs suggestion
+
+No code changes.

package/templates/cursor/skills/shared/verbs/review/orchestrator/contract.yaml

@@ -0,0 +1,25 @@
+name: verb-review-orchestrator
+purpose: Analyze and report on code or docs per review-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded
+  - domain inventory loaded when slug resolves
+  - object skills from graph loaded
+  - no file edits performed
+  - findings cite path:line evidence
+failure_conditions:
+  - code changes without explicit user fix request
+dependencies:
+  skills: [discovery, architecture, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/master_rules.md
+    - docs/architecture.md
+verification: []

package/templates/cursor/skills/shared/verbs/test/orchestrator/SKILL.md

@@ -0,0 +1,24 @@
+---
+name: verb-test-orchestrator
+description: Orchestrates test-* except test-module resolver and test-function. Internal only.
+disable-model-invocation: true
+---
+
+# test-* orchestrator
+
+**Object** from graph. `test-module` → domain resolver. `test-function` → dedicated test-function orchestrator.
+
+Read [_dispatch-utils.md](../_dispatch-utils.md) first.
+
+## Phases
+
+1. **policies** — load `.cursor/policies/`
+2. **load_inventory** — when domain slug maps to inventory
+3. [discovery](../../discovery/SKILL.md)
+4. [planning](../../planning/SKILL.md) — which tests cover intent
+5. **object_skills** — from graph `object_skills.<object>`
+6. [testing](../../testing/SKILL.md); object `schema` → migration/type tests; `integration` → route tests
+7. [verification](../../verification/SKILL.md)
+8. [reporting](../../reporting/SKILL.md)
+
+Fix failures only when intent requests it (then `execution` + governance/implementation).

package/templates/cursor/skills/shared/verbs/test/orchestrator/contract.yaml

@@ -0,0 +1,25 @@
+name: verb-test-orchestrator
+purpose: Verify behavior per test-* command
+inputs:
+  domain:
+    required: false
+  intent:
+    required: true
+outputs:
+  report:
+    type: markdown
+success_criteria:
+  - All policies loaded
+  - domain inventory loaded when slug resolves
+  - relevant tests executed or gap reported
+  - fallow check_changed when files were fixed per intent
+failure_conditions:
+  - claim pass without running tests when tests exist
+dependencies:
+  skills: [discovery, planning, testing, verification, reporting]
+  policies: [master-rules, implementation, mcp-and-deploy]
+  docs:
+    - docs/master_rules.md
+verification:
+  - run_tests
+  - fallow_check_changed