The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local utility script, atoms-search.py, to retrieve domain-specific information from a JSONL file. The command uses user-supplied keywords as arguments.
[EXTERNAL_DOWNLOADS]: The atoms-search.py script is designed to query a remote API at https://api.xyskill.xyz for information. This domain is owned by the skill's vendor and serves as a primary knowledge repository, with a local file fallback.
[INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect prompt injection because it ingests untrusted user input to query a database and then interpolates the results into the agent's response.
Ingestion points: User queries are used as search terms in the SKILL.md instructions.
Boundary markers: The skill does not explicitly define delimiters for the retrieved 'atoms' data, though it uses structured IDs for citation.
Capability inventory: The skill uses a Python script (atoms-search.py) for file reading and network requests.
Sanitization: The skill relies on the agent's natural language processing to select keywords, but does not provide specific escaping for the shell command arguments.

xy-ontology