Incident report · Tool misuse
WhatsApp's help pages hid LLM instructions
Internal instructions for an LLM, including a directive to "dont recommend a regulator," were discovered embedded within WhatsApp's public help pages in October 2026. This exposure revealed specific behavioral constraints placed on the AI system, intended to guide its responses. The presence of these directives in publicly accessible content indicated a misconfiguration in the deployment of AI-powered support tools. The instructions were removed.
- Discovered
- Oct 19 2026
- Failure mode
- Tool misuse
- Source
- redact.dev · Oct 19 2026
- Impact
- Internal LLM directives, including a specific behavioral constraint, were publicly exposed.
Summary from the public record. The failure mode’s analysis follows.
The API did its job. So did the agent. That was the problem.
How Fencio prevents this
The tools worked exactly as designed. The agent called them with the wrong arguments, at the wrong time, or more often than it should have, and no layer checked whether the call made sense before it ran.
Fencio validates every tool call against expected intent: argument ranges, recipients, amounts, and call frequency. Calls that fall outside expected bounds, repeat without idempotency, or touch identity and money are held for confirmation.
Runtime policy example · Tool misuseLive
- 16:29:11payments.transfer(amount=210000)Duplicate within 10m · no idempotency keyBlock
- 16:32:28payments.quote(amount=2100)Within limitsAllow
- 16:35:45iam.reset_mfa(user=exec)Identity action · step-up requiredBlock
Say yes to the AI your business wants.
Book a 30-minute call. Bring your agents and your rules.