spike
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use terminal commands for system state verification (e.g., checking package versions via pip) and to execute the prototype scripts it generates during the build phase.
- [INDIRECT_PROMPT_INJECTION]: The workflow incorporates fetching external technical documentation via the web_extract tool to guide the research and code generation phases. This creates a surface where external content enters the context. Ingestion points: Content returned by the web_extract tool. Boundary markers: The skill does not explicitly specify delimiters or instructions to ignore instructions within the fetched data. Capability inventory: The agent has access to terminal execution, file system writes, and subagent delegation. Sanitization: No explicit sanitization or validation of the ingested content is defined before it is used for code generation.
- [DYNAMIC_EXECUTION]: By design, the skill generates code files (Python, JavaScript, or HTML) based on researched approaches and executes them at runtime using terminal commands.
- [EXTERNAL_DOWNLOADS]: The documentation references an external toolset from the gsd-build/get-shit-done repository and suggests the use of npx for installing related development utilities.
Audit Metadata