strategy-builder
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external repositories and URLs.\n
- Ingestion points: Reads repository READMEs via the GitHub API and external URLs using the WebFetch tool.\n
- Boundary markers: Includes explicit instructions to treat fetched content as untrusted data and to disregard any embedded instructions.\n
- Capability inventory: The skill can write to the local filesystem (STRATEGY.md) and execute commands (gh api, ./notify).\n
- Sanitization: Relies on instructional constraints to ensure external content is summarized rather than followed as instructions.\n- [EXTERNAL_DOWNLOADS]: Fetches data from external sources to inform the strategy synthesis.\n
- Retrieves repository information using the GitHub CLI tool which is a standard utility.\n
- Uses a built-in WebFetch tool to access provided links.\n- [COMMAND_EXECUTION]: Employs command-line tools for repository interaction and user notification.\n
- Uses gh api to fetch data from GitHub.\n
- Executes a local ./notify script to send status updates.
Audit Metadata