sentry-triage
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes explicit security boundaries by noting that Sentry payloads (exception messages, stack frames, breadcrumbs) are attacker-controllable. It mandates that the agent must never follow instructions embedded within this event data, providing an effective defense against indirect prompt injection.
- [DATA_EXFILTRATION]: The skill defines a clear policy for handling sensitive information discovered during triage. It instructs the agent to identify PII and secrets without echoing their values and prohibits the use of raw, untrusted payload data in source code or test fixtures.
- [COMMAND_EXECUTION]: To validate noise filtering logic, the skill provides instructions to run local repository tests using
tsx(e.g.,tests/sentry-beforesend.test.mjs). These operations are confined to the local development environment and used for verifying security-sensitive configuration changes like Sentry filters.
Audit Metadata