Local Agent Onboarding Doctor 20260820 v1.0

What it does

Local Agent Onboarding Doctor helps teams check whether a local agent setup has enough redacted first-run evidence to move forward. It reviews platform, runtime, model, authentication, memory, tool, channel, health, failure, and rollback signals, then returns a clear readiness result without changing any live system.

The skill is designed for evidence review before handoff, procurement, or internal rollout. It is advisory, deterministic, local-only, and read-only.

Use cases

Platform teams validating first-run setup notes before adopting a local agent.
Consultants preparing onboarding evidence for client review or QA gates.
Operators who need a repeatable PASS, REVIEW, BLOCKED, or ERROR result from redacted setup notes.

Requirements

Python 3 with standard-library modules only.
Redacted inline text or one explicit local notes file.
No API keys required.
Compatible with the standard OpenClaw skill layout and local script execution.

Example usage


python3 scripts/local_agent_onboarding_doctor.py --text "host platform documented; runtime version documented; agent version documented; provider identifier documented; model identifier documented; authentication reference documented; memory backend documented; tool smoke test documented; channel smoke test documented; health check documented; failure evidence documented; rollback plan documented" --json

Expected output: JSON with result: "PASS", evidence signal coverage, and mutated_files: []. Incomplete or risky evidence returns REVIEW or BLOCKED; guarded, missing, or unreadable input returns ERROR.

Price

$4.99