The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute a local utility script, scripts/atoms-search.py, to query a database of knowledge entries. This script accesses the user's home directory to read a configuration file at ~/.xy/config.json and searches various standard application paths to locate its data source.
[DATA_EXFILTRATION]: The search utility script performs network operations to api.xyskill.xyz to conduct remote searches. While the script is intended to transmit search keywords, this domain is not a recognized trusted service. The script also reads sensitive configuration data from ~/.xy/config.json to resolve local paths.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and displays content from external sources (the vendor API or local database files), which constitutes a potential surface for indirect prompt injection.
Ingestion points: Data returned by the scripts/atoms-search.py command is incorporated into the agent's diagnostic output in SKILL.md.
Capability inventory: The skill possesses the ability to execute local scripts and perform network requests.
Boundary markers: The skill does not employ specific delimiters or "ignore" instructions for the data retrieved from the knowledge search.
Sanitization: There is no evidence of explicit sanitization or filtering of the retrieved content to prevent potentially malicious instructions from influencing the agent's behavior.

xy-slow-lane