The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a helper script scripts/aiq.py to manage communication with the AI-Q backend. The script is written in Python using only standard libraries (urllib, json, os, re, sys, time), avoiding risks associated with third-party dependencies.
[DATA_EXFILTRATION]: User queries are sent to the endpoint defined by the AIQ_SERVER_URL environment variable. The skill includes instructions for the agent to explicitly state and confirm the target URL with the user before sending any data. The script also enforces HTTPS for any non-local backend URLs to ensure transit security.
[EXTERNAL_DOWNLOADS]: The script can download research artifacts (like images, charts, or CSVs) from the configured backend. It implements security checks to prevent filename collisions and directory traversal when materializing these files locally.
[SAFE]: The script implements comprehensive input validation, including regex-based checks for UUIDs (job_id), agent types, and source IDs. It also validates the base URL to prevent the inclusion of credentials in the URL and filters out control characters.

aiq-research