sentry
Warn
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill metadata incorrectly identifies the author as a trusted organization, which contradicts the actual skill authorship. This discrepancy could lead users to misjudge the skill's safety or official status.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Sentry API, which could contain malicious instructions embedded in error logs or issue descriptions.
- Ingestion points: The
scripts/sentry_api.pyscript retrieves issue and event data from Sentry API GET endpoints. - Boundary markers: No explicit delimiters or instructions are used in the prompt to separate external data from system instructions.
- Capability inventory: The skill utilizes local script execution via
python3and network access for API communication. - Sanitization: The script redacts PII like emails and IP addresses but does not filter for potential prompt injection content.
Audit Metadata