aiox-core 5.0.3 → 5.0.4

package/pro/squads/squad-creator-pro/tasks/detect-squad-context.md

@@ -1,81 +1,81 @@
-# Task: Detect Squad Context
-
-**Task ID:** detect-squad-context
-**Version:** 1.0.0
-**Execution Type:** Hybrid (Script-first detection + Agent interpretation only for ambiguity)
-**Purpose:** Detect the right creation or upgrade route before any squad generation begins
-**Orchestrator:** @squad-chief
-**Mode:** Automatic with conditional elicitation
-
-## Core Principle
-
-```
-Wrong context selection creates wrong architecture.
-Detect first, then build.
-```
-
----
-
-## Inputs
-
-| Parameter | Type | Required | Description |
-|-----------|------|----------|-------------|
-| `domain` | string | Yes | Domain to create/upgrade |
-| `squad_name` | string | No | Existing or desired squad slug |
-| `intent` | enum | No | `create`, `upgrade`, `unsure` |
-
----
-
-## Context Categories
-
-```yaml
-contexts:
-  greenfield_pure:
-    signal: "No existing squad and no prior artifacts"
-    action: "Run greenfield creation workflow"
-
-  pre_existing_brief:
-    signal: "PRD/brief exists but squad folder is missing or partial"
-    action: "Create using brief as source of truth"
-
-  legacy_assets:
-    signal: "Squad exists with assets but missing current standards"
-    action: "Route to brownfield upgrade workflow"
-
-  partial_squad:
-    signal: "Squad exists with incomplete structure"
-    action: "Resume creation with gap fill"
-```
-
----
-
-## Detection Flow
-
-1. Check `squads/{squad_name}/` existence.
-2. Check presence of `config.yaml`, `agents/`, `tasks/`, `workflows/`.
-3. Scan for PRD/brief references in `docs/` and `outputs/`.
-4. Build confidence score per context.
-5. If confidence < 0.70, ask user with 3 options.
-
----
-
-## Outputs
-
-```yaml
-context_detection:
-  detected_context: greenfield_pure | pre_existing_brief | legacy_assets | partial_squad
-  confidence: 0.0-1.0
-  recommended_workflow: wf-context-aware-create-squad | wf-brownfield-upgrade-squad
-  evidence:
-    - file_or_signal: "..."
-      impact: "..."
-  next_action: "..."
-```
-
----
-
-## Validation
-
-- Context assigned with confidence >= 0.70, or user-confirmed.
-- Recommended workflow exists and is executable.
-- Decision evidence saved in logs.
+# Task: Detect Squad Context
+
+**Task ID:** detect-squad-context
+**Version:** 1.0.0
+**Execution Type:** Hybrid (Script-first detection + Agent interpretation only for ambiguity)
+**Purpose:** Detect the right creation or upgrade route before any squad generation begins
+**Orchestrator:** @squad-chief
+**Mode:** Automatic with conditional elicitation
+
+## Core Principle
+
+```
+Wrong context selection creates wrong architecture.
+Detect first, then build.
+```
+
+---
+
+## Inputs
+
+| Parameter | Type | Required | Description |
+|-----------|------|----------|-------------|
+| `domain` | string | Yes | Domain to create/upgrade |
+| `squad_name` | string | No | Existing or desired squad slug |
+| `intent` | enum | No | `create`, `upgrade`, `unsure` |
+
+---
+
+## Context Categories
+
+```yaml
+contexts:
+  greenfield_pure:
+    signal: "No existing squad and no prior artifacts"
+    action: "Run greenfield creation workflow"
+
+  pre_existing_brief:
+    signal: "PRD/brief exists but squad folder is missing or partial"
+    action: "Create using brief as source of truth"
+
+  legacy_assets:
+    signal: "Squad exists with assets but missing current standards"
+    action: "Route to brownfield upgrade workflow"
+
+  partial_squad:
+    signal: "Squad exists with incomplete structure"
+    action: "Resume creation with gap fill"
+```
+
+---
+
+## Detection Flow
+
+1. Check `squads/{squad_name}/` existence.
+2. Check presence of `config.yaml`, `agents/`, `tasks/`, `workflows/`.
+3. Scan for PRD/brief references in `docs/` and `outputs/`.
+4. Build confidence score per context.
+5. If confidence < 0.70, ask user with 3 options.
+
+---
+
+## Outputs
+
+```yaml
+context_detection:
+  detected_context: greenfield_pure | pre_existing_brief | legacy_assets | partial_squad
+  confidence: 0.0-1.0
+  recommended_workflow: wf-context-aware-create-squad | wf-brownfield-upgrade-squad
+  evidence:
+    - file_or_signal: "..."
+      impact: "..."
+  next_action: "..."
+```
+
+---
+
+## Validation
+
+- Context assigned with confidence >= 0.70, or user-confirmed.
+- Recommended workflow exists and is executable.
+- Decision evidence saved in logs.

package/pro/squads/squad-creator-pro/tasks/lookup-model.md

@@ -1,78 +1,78 @@
-# Task: lookup-model
-
-**Command:** `*lookup-model <task-name>`
-**Execution Type:** Worker (deterministic, no LLM needed)
-
-## Purpose
-
-Look up the recommended model tier for a Squad Creator task before execution.
-Returns: `haiku`, `sonnet`, or `opus`.
-
-## Usage
-
-```bash
-# In Squad Chief or any workflow
-model=$(lookup-model "create-agent.md")
-# Returns: opus
-
-model=$(lookup-model "validate-squad.md")
-# Returns: haiku
-```
-
-## Workflow
-
-### Step 1: Load Config
-
-Read `config/model-routing.yaml`
-
-### Step 2: Lookup Task
-
-```yaml
-task_name: "{input}"
-config: model-routing.yaml
-
-lookup:
-  path: tasks.{task_name}.tier
-  default: opus  # Conservative default
-```
-
-### Step 3: Return Result
-
-```yaml
-output:
-  task: "{task_name}"
-  tier: "{haiku|sonnet|opus}"
-  confidence: "{high|medium|low}"
-  reason: "{why this tier}"
-```
-
-## Quick Reference
-
-| Tier | Tasks Count | Use For |
-|------|-------------|---------|
-| **haiku** | 13 | Validation, scoring, admin |
-| **sonnet** | 17 | Documentation, templates, moderate analysis |
-| **opus** | 12 | DNA extraction, agent creation, research |
-
-## Integration with Task Tool
-
-When spawning agents, use the tier as model parameter:
-
-```python
-# Pseudo-code for Squad Chief
-task = "validate-squad.md"
-tier = lookup_model(task)  # Returns "haiku"
-
-Task(
-  subagent_type="general-purpose",
-  model=tier,  # "haiku"
-  prompt="Execute validation..."
-)
-```
-
-## Completion Criteria
-
-- [ ] Task name provided
-- [ ] Config loaded
-- [ ] Tier returned
-- [ ] If task not found, return "opus" (safe default)
+# Task: lookup-model
+
+**Command:** `*lookup-model <task-name>`
+**Execution Type:** Worker (deterministic, no LLM needed)
+
+## Purpose
+
+Look up the recommended model tier for a Squad Creator task before execution.
+Returns: `haiku`, `sonnet`, or `opus`.
+
+## Usage
+
+```bash
+# In Squad Chief or any workflow
+model=$(lookup-model "create-agent.md")
+# Returns: opus
+
+model=$(lookup-model "validate-squad.md")
+# Returns: haiku
+```
+
+## Workflow
+
+### Step 1: Load Config
+
+Read `config/model-routing.yaml`
+
+### Step 2: Lookup Task
+
+```yaml
+task_name: "{input}"
+config: model-routing.yaml
+
+lookup:
+  path: tasks.{task_name}.tier
+  default: opus  # Conservative default
+```
+
+### Step 3: Return Result
+
+```yaml
+output:
+  task: "{task_name}"
+  tier: "{haiku|sonnet|opus}"
+  confidence: "{high|medium|low}"
+  reason: "{why this tier}"
+```
+
+## Quick Reference
+
+| Tier | Tasks Count | Use For |
+|------|-------------|---------|
+| **haiku** | 13 | Validation, scoring, admin |
+| **sonnet** | 17 | Documentation, templates, moderate analysis |
+| **opus** | 12 | DNA extraction, agent creation, research |
+
+## Integration with Task Tool
+
+When spawning agents, use the tier as model parameter:
+
+```python
+# Pseudo-code for Squad Chief
+task = "validate-squad.md"
+tier = lookup_model(task)  # Returns "haiku"
+
+Task(
+  subagent_type="general-purpose",
+  model=tier,  # "haiku"
+  prompt="Execute validation..."
+)
+```
+
+## Completion Criteria
+
+- [ ] Task name provided
+- [ ] Config loaded
+- [ ] Tier returned
+- [ ] If task not found, return "opus" (safe default)