First Hire Process Builder

first-hire-process-builder

Agent: Agent Operations Manager

L2 Agent Operations Manager (1x) responsible for message passing infrastructure, context sharing protocols, inter-agent coordination, and agent health monitoring.

Department ethos: ideal-agent-operations.md

Skill Description

Designs the testing, validation, and deployment procedure for the first production agent, establishing the template for all future agent deployments.

When to Use

• When the organization is deploying its first production agent and no deployment procedure exists.
• When the existing deployment process was ad-hoc and needs to be formalized as a repeatable procedure.
• When a new agent category is being introduced and requires a tailored deployment process distinct from existing ones.

Workflow

1. Define Success Criteria: Specify what "deployed and working" means for the first agent -- accuracy thresholds, latency targets, error rate limits, and integration test pass rates. Deliverable: success criteria document.
2. Design Testing Pipeline: Build the testing sequence: unit tests for individual skills, integration tests for inter-agent communication, and end-to-end tests for complete workflows. Deliverable: test plan with test cases.
3. Create Validation Checklist: Define pre-deployment gates: contract compliance, governance handbook compliance, tool access verification, and security review. Deliverable: deployment validation checklist.
4. Build Deployment Runbook: Write step-by-step deployment instructions including rollback procedures, monitoring setup, and post-deployment smoke tests. Deliverable: deployment runbook.
5. Execute and Document: Deploy the first agent following the runbook. Record deviations, issues encountered, and lessons learned. Update the procedure for future deployments. Deliverable: deployment log and updated procedure.

Anti-Patterns

• Skipping the first-agent rigor: Treating the first deployment as a throwaway prototype that will be "done properly later." Why: the first deployment sets the standard; shortcuts become permanent patterns.
• Testing only happy paths: Validating only expected inputs without testing error handling, edge cases, and failure modes. Why: production traffic includes malformed inputs, timeouts, and cascading failures from day one.
• No rollback plan: Deploying without a documented way to revert if the agent underperforms. Why: production issues demand immediate response; figuring out rollback during an incident wastes critical time.

Output

On success: Produces a deployment procedure package containing success criteria, test plan, validation checklist, and deployment runbook. Delivered as versioned documents ready for execution and reuse.

On failure: Report which procedure components could not be completed (missing infrastructure, undefined success criteria), what was drafted, and what prerequisites must be resolved.

Related Skills

• employment-agreement-setup -- The agent's contract must be defined before deployment validation.
• onboarding-programme-builder -- Post-deployment warm-up follows the deployment procedure.
• employee-handbook-v1 -- Governance policies that the deployment procedure must enforce.