Claude adds policy checks before prompts and tool calls run
Claude Enterprise’s inference hooks are generally available, while its security center and centralized management remain beta and sign-out does not revoke running cloud work or API keys.
Anthropic’s November 5 enterprise update makes prompt hooks and pre-tool-call hooks generally available, connecting supported Claude surfaces to an organization’s security tools before the relevant activity proceeds.
The security center, which compares configuration with Anthropic’s recommendations, remains beta. Its report distinguishes automatically checked settings from administrator attestations and is not a compliance certification. [1][2][3]
The beta member sign-out API ends sign-in sessions across devices, but does not terminate API keys, already-running cloud sessions or scheduled routines. A person can also sign in again. That makes sign-out useful for a session reset, but insufficient as a complete offboarding or incident-containment procedure. The response must address the remaining credentials and workloads separately. [1][6]
Sources & details 6 sources
Discuss in your voice app
Take the reporting, visual explanations and original sources into a conversation. Copy the brief, paste it into your preferred assistant, then start its voice mode.
Paste the brief into your voice app to discuss this reporting.Reporting sources
- Anthropic — Enterprise security update
November 5 release states GA versus beta, sign-out exclusions and detection-only local EDR.
- Claude Help — inference hooks
Supported prompt and pre-tool-call policy integration; provider documentation, not independent enforcement testing.
- Claude Help — security center
Beta, view-only settings assessment with attestations; expressly not certification.
- Claude Help — EDR integration
Beta local-VM coverage, third-party sensor telemetry and cloud exclusion.
- Claude Help — centralized management
Enterprise beta, eligibility limitations and irreversible enrollment.
- Claude Help — member management
Member administration and sign-out scope; operational corroboration from the same provider.
Images & figures
Inference hooks connect prompts and pre-tool calls to an organization’s policy checks before supported activity proceeds. The separate security-center report is a configuration assessment, not certification.
Euangelion · Source · Usage terms