The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data and user input, creating a surface for potential injection attacks. * Ingestion points: Research topics and goals provided by the user in SKILL.md, and raw content from multiple web searches performed in step-02 through step-06. * Boundary markers: Absent. The instructions do not define specific delimiters or instructions to ignore embedded commands in external data. * Capability inventory: The skill can execute local scripts via uv run and write to the local file system to generate reports. * Sanitization: No explicit sanitization or filtering of web search results or user input is implemented.
[COMMAND_EXECUTION]: The skill uses uv run to execute local Python scripts like resolve_customization.py. These scripts are part of the workflow for resolving configuration settings and executing post-completion actions.
[DYNAMIC_EXECUTION]: The agent dynamically loads configuration from multiple levels of TOML files and can load facts from the local file system using glob patterns. This flexible setup is used to tailor the research workflow to specific project needs.

gds-domain-research