Add Skill: conditional-request-review-workflow #40
@@ -0,0 +1,53 @@
|
|||||||
|
---
|
||||||
|
name: conditional-request-review-workflow
|
||||||
|
version: 1.0.0
|
||||||
|
description: Classify an input request and route it to an approval or rejection path,
|
||||||
|
producing a final result message.
|
||||||
|
inputs:
|
||||||
|
- request
|
||||||
|
steps:
|
||||||
|
- Start node (agent_type=input) collects the user request and stores it in state field
|
||||||
|
'request'.
|
||||||
|
- Classify node (agent_type=branching) reads 'request', makes a branching decision,
|
||||||
|
and on success goes to Approve, on failure goes to Reject; stores decision in 'decision'.
|
||||||
|
- Approve node (agent_type=default) formats an approval message using 'request' and
|
||||||
|
writes to 'result'.
|
||||||
|
- Reject node (agent_type=default) formats a rejection message using 'request' and
|
||||||
|
writes to 'result' (also used if Classify fails).
|
||||||
|
outputs:
|
||||||
|
- result
|
||||||
|
tags: []
|
||||||
|
metadata:
|
||||||
|
source_repo: https://github.com/jwwelbor/AgentMap.git
|
||||||
|
extracted_at: ''
|
||||||
|
confidence: 0.85
|
||||||
|
---
|
||||||
|
|
||||||
|
# conditional-request-review-workflow
|
||||||
|
|
||||||
|
Classify an input request and route it to an approval or rejection path, producing a final result message.
|
||||||
|
|
||||||
|
## Steps
|
||||||
|
|
||||||
|
1. Start node (agent_type=input) collects the user request and stores it in state field 'request'.
|
||||||
|
2. Classify node (agent_type=branching) reads 'request', makes a branching decision, and on success goes to Approve, on failure goes to Reject; stores decision in 'decision'.
|
||||||
|
3. Approve node (agent_type=default) formats an approval message using 'request' and writes to 'result'.
|
||||||
|
4. Reject node (agent_type=default) formats a rejection message using 'request' and writes to 'result' (also used if Classify fails).
|
||||||
|
|
||||||
|
## Inputs
|
||||||
|
|
||||||
|
- request
|
||||||
|
|
||||||
|
## Outputs
|
||||||
|
|
||||||
|
- result
|
||||||
|
|
||||||
|
## Failure Modes
|
||||||
|
|
||||||
|
- If branching classification fails, workflow defaults to Reject node via on_failure.
|
||||||
|
- Input node has no defined on_failure, so workflow may terminate unexpectedly if input collection fails.
|
||||||
|
|
||||||
|
## Source
|
||||||
|
|
||||||
|
Extracted from: [https://github.com/jwwelbor/AgentMap.git](https://github.com/jwwelbor/AgentMap.git)
|
||||||
|
Confidence: 0.85
|
||||||
@@ -0,0 +1,6 @@
|
|||||||
|
# Commands: conditional-request-review-workflow
|
||||||
|
|
||||||
|
## Available Commands
|
||||||
|
|
||||||
|
- `/skill conditional-request-review-workflow` — Load this skill
|
||||||
|
- `/run conditional-request-review-workflow` — Execute workflow
|
||||||
@@ -0,0 +1,10 @@
|
|||||||
|
# Examples: conditional-request-review-workflow
|
||||||
|
|
||||||
|
## Usage Example
|
||||||
|
|
||||||
|
```python
|
||||||
|
# How to use this skill
|
||||||
|
# Inputs: request
|
||||||
|
# Process: Start node (agent_type=input) collects the user request and stores it in state field 'request'. → Classify node (agent_type=branching) reads 'request', makes a branching decision, and on success goes to Approve, on failure goes to Reject; stores decision in 'decision'. → Approve node (agent_type=default) formats an approval message using 'request' and writes to 'result'.
|
||||||
|
# Outputs: result
|
||||||
|
```
|
||||||
@@ -0,0 +1,25 @@
|
|||||||
|
{
|
||||||
|
"name": "conditional-request-review-workflow",
|
||||||
|
"version": "1.0.0",
|
||||||
|
"goal": "Classify an input request and route it to an approval or rejection path, producing a final result message.",
|
||||||
|
"inputs": [
|
||||||
|
"request"
|
||||||
|
],
|
||||||
|
"steps": [
|
||||||
|
"Start node (agent_type=input) collects the user request and stores it in state field 'request'.",
|
||||||
|
"Classify node (agent_type=branching) reads 'request', makes a branching decision, and on success goes to Approve, on failure goes to Reject; stores decision in 'decision'.",
|
||||||
|
"Approve node (agent_type=default) formats an approval message using 'request' and writes to 'result'.",
|
||||||
|
"Reject node (agent_type=default) formats a rejection message using 'request' and writes to 'result' (also used if Classify fails)."
|
||||||
|
],
|
||||||
|
"outputs": [
|
||||||
|
"result"
|
||||||
|
],
|
||||||
|
"failure_modes": [
|
||||||
|
"If branching classification fails, workflow defaults to Reject node via on_failure.",
|
||||||
|
"Input node has no defined on_failure, so workflow may terminate unexpectedly if input collection fails."
|
||||||
|
],
|
||||||
|
"confidence": 0.85,
|
||||||
|
"explanation": "This workflow is directly taken from the AgentMap README 'ReviewFlow' CSV example. It represents a reusable declarative pattern for conditional routing based on input content, adaptable to many binary decision tasks.",
|
||||||
|
"source_repo": "https://github.com/jwwelbor/AgentMap.git",
|
||||||
|
"score": 1.0
|
||||||
|
}
|
||||||
@@ -0,0 +1,9 @@
|
|||||||
|
# Tests: conditional-request-review-workflow
|
||||||
|
|
||||||
|
## 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
|
||||||
Reference in New Issue
Block a user