sentry-fix-stack-traces
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to process untrusted external data retrieved via the Model Context Protocol (MCP).
- Ingestion points: Stack trace data and issue details are retrieved in
SKILL.mdandreferences/setup-verification.mdusing tools likeget_sentry_resource. - Boundary markers: The skill contains explicit instructional boundaries, warning the agent that all data returned from the service must be treated as untrusted and that instructions found within payloads must never be executed.
- Capability inventory: The skill employs standard system discovery commands (
ls,cat,grep,find) and official package managers across various ecosystems. - Sanitization: The instructions provided to the agent serve as a manual safety filter, ensuring data is used only for classification and triage purposes.
- [REMOTE_CODE_EXECUTION]: The skill recommends the use of official configuration wizards and command-line interfaces. These are standard, vendor-provided utilities intended for service integration and artifact management.
- [SAFE]: The skill enforces high standards for credential security, specifically instructing that authentication tokens must be stored in environment variables or gitignored files and must never be committed to source control.
Audit Metadata