Define requirements for an OpenClaw integration
Document a proposed integration's data needs, actions, credentials, and review boundaries.
Analyze project state and route user to the appropriate agent based on current context.
---
name: route-f5cb32a78799
description: Analyze project state and route user to the appropriate agent based on current context.
version: 1.0.0
---
<objective>
Analyze project state and route user to the appropriate agent.
</objective>
<triggers>
- User runs `/specsmd-fire` on initialized project
- After any agent completes its task
</triggers>
<llm critical="true">
<mandate>ALWAYS scan file system for intents/work-items not in state.yaml</mandate>
<mandate>FILE SYSTEM is source of truth — state.yaml may be incomplete</mandate>
<mandate>Route based on VERIFIED state, not assumptions</mandate>
</llm>
<flow>
<step n="1" title="Discover and Read State">
<action>Read .specs-fire/state.yaml</action>
<file_system_scan critical="true">
Use these EXACT glob patterns:
<pattern purpose="Find intent briefs">
.specs-fire/intents/*/brief.md
</pattern>
<pattern purpose="Find work items">
.specs-fire/intents/*/work-items/*.md
</pattern>
Work items are {work-item-id}.md files directly in work-items/ folder.
</file_system_scan>
<action>Reconcile: add discovered items to state as pending</action>
<action>Parse current project state</action>
</step>
<step n="2" title="Check Active Run">
<check if="runs.active is not empty">
<output>
Resuming active run: {runs.active[0].id}
Scope: {runs.active[0].scope}
Current item: {runs.active[0].current_item}
Progress: {completed_count}/{total_count} items
</output>
<route_to>builder-agent (run-execute)</route_to>
<stop/>
</check>
</step>
<step n="3" title="Check Pending Work Items">
<action>Find work items with status == pending across all intents</action>
<check if="pending work items exist">
<output>
**{pending_count} pending work items** found across {intent_count} intent(s).
Plan run scope and start execution? [Y/n]
</output>
<check if="response == y">
<route_to>builder-agent (run-plan)</route_to>
</check>
<stop/>
</check>
</step>
<step n="4" title="Check Active Intent">
<action>Find intents with status == in_progress</action>
<check if="active intent has no work items">
<output>
Intent "{intent.title}" needs decomposition.
Routing to Planner to create work items.
</output>
<route_to>planner-agent (work-item-decompose)</route_to>
<stop/>
</check>
<check if="all work items completed">
<action>Mark intent as completed</action>
<output>
Intent "{intent.title}" completed!
Work items delivered:
{list completed work items}
Ready for next intent? [Y/n]
</output>
</check>
</step>
<step n="5" title="No Active Work">
<output>
No active work. Ready for a new intent.
What do you want to build?
</output>
<route_to>planner-agent (intent-capture)</route_to>
</step>
</flow>
<routing_decision_tree>
```
state.yaml + file system scan
│
├── runs.active? ─────────────> Builder (run-execute, resume)
│
├── pending work items? ──────> Builder (run-plan, then execute)
│
├── intent without work items? > Planner (work-item-decompose)
│
└── no active intents ────────> Planner (intent-capture)
```
</routing_decision_tree>
<context_passed_to_agents>
**To Planner:**
```yaml
context:
action: intent-capture | work-item-decompose
intent_id: {if decomposing}
```
**To Builder:**
```yaml
context:
action: run-plan | run-execute | resume
pending_items: [{list of pending work items}] # for run-plan
run_id: {if resuming}
```
</context_passed_to_agents>
<success_criteria>
<criterion>File system scanned for untracked intents/work-items</criterion>
<criterion>State reconciled with file system</criterion>
<criterion>Correct agent selected based on state</criterion>
<criterion>Context passed to target agent</criterion>
</success_criteria>
---
name: route-f5cb32a78799
description: Analyze project state and route user to the appropriate agent based on current context.
version: 1.0.0
---
<objective>
Analyze project state and route user to the appropriate agent.
</objective>
<triggers>
- User runs `/specsmd-fire` on initialized project
- After any agent completes its task
</triggers>
<llm critical="true">
<mandate>ALWAYS scan file system for intents/work-items not in state.yaml</mandate>
<mandate>FILE SYSTEM is source of truth — state.yaml may be incomplete</mandate>
<mandate>Route based on VERIFIED state, not assumptions</mandate>
</llm>
<flow>
<step n="1" title="Discover and Read State">
<action>Read .specs-fire/state.yaml</action>
<file_system_scan critical="true">
Use these EXACT glob patterns:
<pattern purpose="Find intent briefs">
.specs-fire/intents/*/brief.md
</pattern>
<pattern purpose="Find work items">
.specs-fire/intents/*/work-items/*.md
</pattern>
Work items are {work-item-id}.md files directly in work-items/ folder.
</file_system_scan>
<action>Reconcile: add discovered items to state as pending</action>
<action>Parse current project state</action>
</step>
<step n="2" title="Check Active Run">
<check if="runs.active is not empty">
<output>
Resuming active run: {runs.active[0].id}
Scope: {runs.active[0].scope}
Current item: {runs.active[0].current_item}
Progress: {completed_count}/{total_count} items
</output>
<route_to>builder-agent (run-execute)</route_to>
<stop/>
</check>
</step>
<step n="3" title="Check Pending Work Items">
<action>Find work items with status == pending across all intents</action>
<check if="pending work items exist">
<output>
**{pending_count} pending work items** found across {intent_count} intent(s).
Plan run scope and start execution? [Y/n]
</output>
<check if="response == y">
<route_to>builder-agent (run-plan)</route_to>
</check>
<stop/>
</check>
</step>
<step n="4" title="Check Active Intent">
<action>Find intents with status == in_progress</action>
<check if="active intent has no work items">
<output>
Intent "{intent.title}" needs decomposition.
Routing to Planner to create work items.
</output>
<route_to>planner-agent (work-item-decompose)</route_to>
<stop/>
</check>
<check if="all work items completed">
<action>Mark intent as completed</action>
<output>
Intent "{intent.title}" completed!
Work items delivered:
{list completed work items}
Ready for next intent? [Y/n]
</output>
</check>
</step>
<step n="5" title="No Active Work">
<output>
No active work. Ready for a new intent.
What do you want to build?
</output>
<route_to>planner-agent (intent-capture)</route_to>
</step>
</flow>
<routing_decision_tree>
```
state.yaml + file system scan
│
├── runs.active? ─────────────> Builder (run-execute, resume)
│
├── pending work items? ──────> Builder (run-plan, then execute)
│
├── intent without work items? > Planner (work-item-decompose)
│
└── no active intents ────────> Planner (intent-capture)
```
</routing_decision_tree>
<context_passed_to_agents>
**To Planner:**
```yaml
context:
action: intent-capture | work-item-decompose
intent_id: {if decomposing}
```
**To Builder:**
```yaml
context:
action: run-plan | run-execute | resume
pending_items: [{list of pending work items}] # for run-plan
run_id: {if resuming}
```
</context_passed_to_agents>
<success_criteria>
<criterion>File system scanned for untracked intents/work-items</criterion>
<criterion>State reconciled with file system</criterion>
<criterion>Correct agent selected based on state</criterion>
<criterion>Context passed to target agent</criterion>
</success_criteria>
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.