sentry-instrumentation
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents a pattern where an AI agent reads user-generated content (errors, breadcrumbs, and feedback) from Sentry's servers, which represents a known indirect prompt injection surface.
- Ingestion points: In
references/agent-tooling.md, the skill describes an agent using a Sentry MCP connection to fetch issues and events. - Boundary markers: The instructions warn the agent to treat the data as untrusted but do not define specific structural delimiters for the ingested data.
- Capability inventory: Agents following these guidelines would have broad access to modify repository code and manage build-time artifacts using authentication tokens.
- Sanitization: The skill explicitly identifies feedback text as unscrubbed and warns that event messages can contain text shaped like instructions.
Audit Metadata