Agently TriggerFlow Playbook

This skill is the scenario-routing entry point for TriggerFlow work in Agently. Use it when the request starts from business workflow requirements rather than specific TriggerFlow APIs. It helps choose the right implementation skill or skill combination. It does not replace the implementation skills themselves.

Prerequisite: Agently >= 4.0.8.5.

Scope

Use this skill for:

• deciding whether a business requirement should be implemented with TriggerFlow
• mapping a workflow requirement to the correct TriggerFlow skill
• deciding when TriggerFlow should be combined with agently-model-setup or agently-output-control
• selecting between general orchestration work, workflow-pattern work, state-and-resource work, sub-flow work, model-integration work, config work, execution-state work, and explicit interrupt or runtime-stream work

Do not use this skill for:

• direct API-level TriggerFlow implementation details
• provider setup or model-request configuration details
• standalone model output parsing without a TriggerFlow workflow

Workflow

1. Start with references/scenario-router.md to map the business requirement to the right capability area.
2. Read references/current-skill-map.md to choose the implementation skill or skill combination.
3. Switch to the selected implementation skill and do the actual coding there.

Routing Rules

• signal-driven workflow primitives, event routing, execution entrypoints, or result semantics -> agently-triggerflow-orchestration
• router patterns, fan-out and fan-in, item-wise worker flows, safe loops, evaluator-optimizer, ReAct loops, or approval-gate workflow design -> agently-triggerflow-patterns
• state placement, flow-versus-execution sharing, runtime-resource boundaries, or restart-safe state design -> agently-triggerflow-state-and-resources
• isolated child workflow, explicit parent-child handoff, or to_sub_flow(...) boundary -> agently-triggerflow-subflows
• workflow step sends model requests, fans out model calls, or uses delta / instant inside the flow -> agently-triggerflow-model-integration
• specialized agent teams, planner-worker, reviewer-reviser, or parallel experts as the main workflow concern -> agently-multi-agent-patterns
• flow definition should be exported, imported, copied, or visualized -> agently-triggerflow-config
• a running execution should be saved, restored, resumed, or rehydrated with resources -> agently-triggerflow-execution-state
• a suspended workflow or conversation should be resumed after restart -> combine agently-triggerflow-config + agently-triggerflow-execution-state, and add agently-session-memo if memory lives outside the execution
• human approval, external resume, waiting execution, or live runtime stream -> agently-triggerflow-interrupts-and-stream
• workflow step needs provider setup -> combine with agently-model-setup
• workflow step needs detailed output-schema or structured-streaming semantics -> combine with agently-output-control

References

• references/scenario-router.md
• references/current-skill-map.md