The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill features an indirect prompt injection surface as it processes user input to drive search queries.
Ingestion points: User-provided product descriptions and market details in Phase 1 and Phase 2 are used as search keywords.
Boundary markers: The instructions do not define clear delimiters or "ignore instructions" wrappers for the content retrieved from the search tool.
Capability inventory: The skill has the ability to execute the bundled atoms-search.py script and perform network requests to the vendor's API.
Sanitization: The Python search script utilizes argparse for parameter parsing, but there is no explicit filtering or sanitization of the knowledge data retrieved before it is returned to the agent's context.
[COMMAND_EXECUTION]: The skill instructs the agent to execute a locally bundled Python script (scripts/atoms-search.py) to retrieve evidence and case studies relevant to the user's query.
[EXTERNAL_DOWNLOADS]: The search script performs network requests to https://api.xyskill.xyz. This is a vendor-owned domain used to fetch supplementary knowledge atoms, which is consistent with the skill's purpose and the author's identity.

xy-selection