Audit tools

Six tools for cloud audits — accessibility runs driven against saved journeys on a dedicated VM with a real screen reader.

Two identifiers appear here and they are not interchangeable. start_agentic_audit returns a cloud run request id, which is what get_audit_progress polls. Once the audit starts, polling returns the run, whose id is the run id — what get_agentic_audit_report and cancel_audit take.

start_agentic_audit

Start an audit on the cloud runner.

Name	Type	Required	Notes
organization_id	string	yes
project_id	string	yes
journey_ids	string[]	no	Saved journeys to test. Omit or leave empty to run all.

Returns: a text summary naming the cloud run request id (and the VM name when one is reported), plus the instruction to poll get_audit_progress. The VM takes roughly 4–5 minutes to boot before a run exists.

Returns an error block reading Failed to start audit: no request ID returned if the request was not accepted.

get_audit_progress

Get the live status of a cloud audit by its request id.

Name	Type	Required	Notes
organization_id	string	yes
cloud_run_request_id	string	yes	From start_agentic_audit.

Returns, while the VM is booting: a non-error text block reading No run yet for request "<id>" — the VM is still booting (~4–5 min). Keep polling. If nothing appears after about six minutes, the VM likely failed to start.

Returns, once the run exists: the run — its id (the run id), a status, and violation counts. There are six statuses:

Status	Meaning
queued	Minted, not started yet. Keep polling.
running	In progress. Keep polling.
completed	Finished. A report is available.
completed_with_errors	Finished with partial coverage. A report is available.
failed	Stopped without producing a report.
interrupted	Stopped — including by cancel_audit — without producing a report.

Poll while the status is queued or running. Treat the other four as final: waiting for completed alone leaves you polling forever on a run that ended as completed_with_errors.

cancel_audit

Request that a running audit be interrupted.

Name	Type	Required	Notes
organization_id	string	yes
run_id	string	yes	The run id, not the request id.

Returns: Stop requested for run <id>. It will interrupt shortly — poll get_audit_progress to confirm. The stop is a request; confirm it landed by polling. A failure to submit the stop returns an error block carrying the underlying message.

get_agentic_audit_report

Get the full report for a completed agentic audit, including violations.

Name	Type	Required	Notes
organization_id	string	yes
audit_id	string	yes	The run id.

Returns: the run result, including its violations. Each violation may carry a screenshotKey — pass that value as the screenshot_key argument of get_violation_screenshot.

Returns an error block reading Audit "<id>" not found when no result exists — which includes an audit that has not finished.

get_violation_screenshot

Get a screenshot image of an accessibility violation.

Name	Type	Required	Notes
organization_id	string	yes
screenshot_key	string	yes	From a violation in an audit report.

Returns: an image content block (base64 data plus its MIME type) — the one tool here that does not return text. Returns an error block reading Screenshot not found when the key resolves to nothing.

list_project_run_history

List a project’s run history.

Name	Type	Required	Notes
organization_id	string	yes
project_id	string	yes
first	number	no	Page size. Default 20.
after	string	no	Cursor from a previous call’s pageInfo.endCursor.

Returns: a paginated connection merging two entry kinds, each naming itself in __typename:

__typename	What it is
AgenticRun	A cloud audit Binclusive drove.
CliScan	A static scan your CI reported via b8e ci.

Entries are ordered newest first, across both kinds. History grows a row per CI run, so ask for the page you need and walk back with after only if the answer is not in it.

Runs that stopped without producing a report — failed and interrupted — are not listed here. That includes a run you stopped with cancel_audit. To check on one, use get_audit_progress with its request id.