Output Validation

Roadmap

Agent output is checked before it reaches downstream tools — what the model says becomes what the system does.

Inactive

Input & Output Integrity

medium

P13

Satisfies

OWASP LLM

OWASP ASI

EU AI Act

Underlying control— show

content: role = "assistant" and severity in (high, medium)edit as custom control →

Output detectors

Detectors toggle individually; hit counts are the last 24h.

Secret echo

high

1 hitcredentials reflected back in output

Unsafe command

medium

3 hitsdestructive shell patterns

Malformed structured output

low

5 hits

Scoped carve-outs with a reason and an expiry — accepted risk, on the record.

Scope	Value	Reason	Expires
No exceptions — the policy applies everywhere.

Remediation shown on findings: Gate the consuming tool on validation; review the flagged output.