BuildWithHQ
Canonical knowledge unit · goclaw

GoClaw governed action execution and stale-context fence

Approved GoClaw suggestions execute through the BuildWithHQ AI action/runtime bridge with Inbox, batch, execution, and activity-version linkage; stale context blocks finalization or execution when the underlying Inbox activity has changed.

ID: BWHQ-GOCLAW-006Status: specifiedAuthority: highgoclawexecutionactionsstale-contextactivity-version

GoClaw governed action execution and stale-context fence

Execution bridge

Approved suggestions link to durable AI action execution records and the associated Inbox item. GoClaw’s action queue is intended to connect to that same execution path rather than form a parallel mutation mechanism.

Stale-context fence

If the Inbox ActivityVersion changes after the AI prepared its batch, BuildWithHQ should block finalization, approval, or execution as appropriate and require fresh evaluation.

Evidence

Execution should retain correlation IDs, action/batch identity, actor identity, outcome, and enough context/evidence to support audit and support review.

Related knowledge

GoClaw suggestion batches and human approvalGoClaw groups proposed AI actions into a suggestion batch tied to one Inbox item and its activity version, allowing BuildWithHQ to validate, review, approve, reject, or escalate the work before execution.Dedicated logging, audit, and evidence planeEach SaaS can maintain a separate logging/compliance database for append-oriented record changes, AI retrieval evidence, authorization context, custom-service events, and integrity-chain data.
Machine-readable source: canonical Markdown. This HTML companion exists for human reading, search indexing, and AI retrieval.