sentry-sdk-skill-creator
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [COMMAND_EXECUTION]: Uses common shell utilities like
ls,grep, andcatto examine project structures and manifest files (package.json,go.mod). This helps the agent recommend the correct Sentry SDKs and framework integrations. - [EXTERNAL_DOWNLOADS]: Promotes the use of
@sentry/wizard, an official Sentry command-line tool. This tool facilitates secure authentication and automated configuration for developers. - [DYNAMIC_EXECUTION]: Implements a systematic research process that uses agent tools to gather technical information from official Sentry documentation and store it in local files for processing.
- [INDIRECT_PROMPT_INJECTION]: The skill's detection phase reads project configuration files which are external to the skill and could contain data designed to influence the agent.
- Ingestion points: Local manifest files (
package.json,go.mod) are read to determine project characteristics. - Boundary markers: There are no explicit delimiters defined for separating project file content from the agent's internal instructions.
- Capability inventory: The skill utilizes file reading, writing, and automated documentation fetching.
- Sanitization: The skill relies on its structured wizard flow rather than formal sanitization of the project metadata it ingests.
- [NO_CODE]: The skill bundle contains only documentation and instructional markdown files without any standalone executable scripts or binaries.
Audit Metadata