skills/conditional-request-review/SKILL.md

@@ -0,0 +1,52 @@
+---
+name: conditional-request-review
+version: 1.0.0
+description: Capture a user request, classify it via branching, and produce an approval
+  or rejection result.
+inputs:
+- request
+steps:
+- 'Start: input agent prompts user for request and stores it in state field ''request'''
+- 'Classify: branching agent evaluates ''request'' and routes to Approve on success
+  or Reject on failure, storing ''decision'''
+- 'Approve: default agent formats approval message using ''request'' and stores it
+  in ''result'''
+- 'Reject: default agent formats rejection message using ''request'' and stores it
+  in ''result'''
+outputs:
+- result (string message indicating approval or rejection)
+tags: []
+metadata:
+  source_repo: https://github.com/jwwelbor/AgentMap.git
+  extracted_at: ''
+  confidence: 0.85
+---
+
+# conditional-request-review
+
+Capture a user request, classify it via branching, and produce an approval or rejection result.
+
+## Steps
+
+1. Start: input agent prompts user for request and stores it in state field 'request'
+2. Classify: branching agent evaluates 'request' and routes to Approve on success or Reject on failure, storing 'decision'
+3. Approve: default agent formats approval message using 'request' and stores it in 'result'
+4. Reject: default agent formats rejection message using 'request' and stores it in 'result'
+
+## Inputs
+
+- request
+
+## Outputs
+
+- result (string message indicating approval or rejection)
+
+## Failure Modes
+
+- If branching classification fails, workflow routes to Reject (on_failure)
+- If input not provided, workflow may hang or error depending on runtime
+
+## Source
+
+Extracted from: [https://github.com/jwwelbor/AgentMap.git](https://github.com/jwwelbor/AgentMap.git)
+Confidence: 0.85

skills/conditional-request-review/commands.md

@@ -0,0 +1,6 @@
+# Commands: conditional-request-review
+
+## Available Commands
+
+- `/skill conditional-request-review` — Load this skill
+- `/run conditional-request-review` — Execute workflow

skills/conditional-request-review/examples.md

@@ -0,0 +1,10 @@
+# Examples: conditional-request-review
+
+## Usage Example
+
+```python
+# How to use this skill
+# Inputs: request
+# Process: Start: input agent prompts user for request and stores it in state field 'request' → Classify: branching agent evaluates 'request' and routes to Approve on success or Reject on failure, storing 'decision' → Approve: default agent formats approval message using 'request' and stores it in 'result'
+# Outputs: result (string message indicating approval or rejection)
+```

skills/conditional-request-review/metadata.json

@@ -0,0 +1,25 @@
+{
+  "name": "conditional-request-review",
+  "version": "1.0.0",
+  "goal": "Capture a user request, classify it via branching, and produce an approval or rejection result.",
+  "inputs": [
+    "request"
+  ],
+  "steps": [
+    "Start: input agent prompts user for request and stores it in state field 'request'",
+    "Classify: branching agent evaluates 'request' and routes to Approve on success or Reject on failure, storing 'decision'",
+    "Approve: default agent formats approval message using 'request' and stores it in 'result'",
+    "Reject: default agent formats rejection message using 'request' and stores it in 'result'"
+  ],
+  "outputs": [
+    "result (string message indicating approval or rejection)"
+  ],
+  "failure_modes": [
+    "If branching classification fails, workflow routes to Reject (on_failure)",
+    "If input not provided, workflow may hang or error depending on runtime"
+  ],
+  "confidence": 0.85,
+  "explanation": "Workflow extracted from AgentMap README example 'ReviewFlow' CSV. It is a generic conditional routing pattern usable for any binary decision process.",
+  "source_repo": "https://github.com/jwwelbor/AgentMap.git",
+  "score": 1.0
+}

skills/conditional-request-review/tests.md

@@ -0,0 +1,9 @@
+# Tests: conditional-request-review
+
+## Test Checklist
+
+- [ ] Workflow has at least 3 steps
+- [ ] All inputs are defined
+- [ ] All outputs are defined
+- [ ] Failure modes are documented
+- [ ] Skill can be loaded without errors