The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Socket CLI through npx socket and curl for API requests. These commands are standard operations for the tool's security research functionality.\n- [EXTERNAL_DOWNLOADS]: Data is retrieved from official vendor endpoints including api.socket.dev and socket.dev. These downloads are necessary for fetching package scores, alerts, and CVE information.\n- [CREDENTIALS_UNSAFE]: Includes a hardcoded public demo token (sktsec_t_--RAN5U4ivauy4w37-6aoKyYPDt5ZbaT5JBVMqiwKo_api). This is a documented vendor credential provided for public trial use and has limited permissions.\n- [PROMPT_INJECTION]: The skill processes untrusted package data from external registries, creating a surface for indirect prompt injection.\n
Ingestion points: API responses from api.socket.dev and web content scraped from socket.dev package pages.\n
Boundary markers: Not explicitly defined in the prompt instructions for isolating external package metadata.\n
Capability inventory: Shell execution via npx socket and network access via curl.\n
Sanitization: No specific sanitization or filtering of external package descriptions is mentioned.

socket-inspect