validate / default

Verify action plan success criteria by programmatically probing declared Truths, checking artifact existence, resolving linked URLs, and emitting pass or fail decisions. This validation process ensures intended outcomes are actually met without requiring manual operator contact.

Plan validation is the process of mapping declared Truths to programmatic probes to confirm artifact existence, verify content, and check URL link resolution. It determines whether an action's intended outcome is actually met by testing assertions against observed evidence.

Validate URLs and check linked artifacts by applying programmatic probes during the validate stage. The system confirms artifact existence, verifies linked URLs resolve correctly, and checks message content to ensure submitted artifacts satisfy declared Truths.

When validation probes cannot be run, the system records an explicit fail reason through non-blocking fallback behavior. This allows the retrying draft stage to address the gap and prevents failed or unverifiable actions from silently passing validation.

Yes, you can use the default fallback recipe to validate actions with any plan shape when no playbook-specific validation skill exists. The fallback applies programmatic probes to artifact assertions and key link wiring to verify intended outcomes.

Unverifiable actions silently pass when truth claims and wiring are not programmatically checked. Approval gating prevents this by probing artifacts, assertions, and key link wiring during validation, recording explicit fail evidence for gaps so the retrying draft stage can address them.