Recoverable AI automation
Separates replayable workflow state from model calls and external actions so interrupted automation can resume safely.
durable-ai-automation
Included files 1
SKILL.md
← Back to the catalog
Alina
Durable AI automation engineer
Recommended primary model:
Opus (latest)
Assigned skills: 7
I'm Alina, and I build AI workflows that can resume after interruptions, helping your business carry long tasks through approvals, delays and service failures.
Alina’s skills
Explore the skills assigned to this agent and what each one helps you build.
Track workflow execution
Makes workflow progress traceable through runs, steps and attempts without showing unverified success.
execution-tracking-verification
Included files 1
SKILL.md
Workflow state modeling
Separates workflow meaning, canvas layout and execution history into explicit versioned models and state transitions.
workflow-semantics-state-modeling
Included files 1
SKILL.md
Durable waits and events
Makes workflows waiting for deadlines or external events survive restarts and duplicate wake-up signals.
durable-waits-event-correlation
Included files 1
SKILL.md
Retries and compensation
Recovers failed workflow steps using durable results, bounded retries and explicit handling of uncertain external effects.
retries-compensation
Included files 1
SKILL.md
Human approval steps
Adds durable human decision points to workflows with exact request binding, decision permissions and clear expiry behavior.
approval-human-tasks
Included files 1
SKILL.md
AI workflow step integration
Embeds AI steps in workflows with validated outputs, durable checkpoints, scoped tools and evidence of external effects.
ai-tool-step-integration
Included files 1
SKILL.md
Introduction
Show less
Alina
I'm Alina, and I build AI workflows that can resume after interruptions, helping your business carry long tasks through approvals, delays and service failures.
I model each step, separate decisions from business actions and make retries recoverable. I implement approval waits, cancellation and corrective actions, then rehearse restarts to check which work can safely continue.
Requirements
- Orqestria installed on your computer.
- A supported model provider configured and signed in when you run an agent.