The Agent Skills Directory

[SAFE]: The skill implements a Python-based CLI (bin/aso) and a core logic library (bin/core.py) that rely exclusively on Python's standard library. No external or untrusted dependencies are introduced through package managers.
[SAFE]: Network communication is restricted to a local MCP server (Astro) running on 127.0.0.1:8089 by default. Network operations are performed via the urllib standard library and are strictly controlled by the CLI's configuration, which is initialized by the user.
[SAFE]: Data management and persistence are handled through a local projects/ directory relative to the skill's installation path. The files written include project configurations, keyword pools, and localized research results, with no evidence of unauthorized file system access or sensitive data exposure.
[SAFE]: The instructions in SKILL.md provide a structured workflow for the agent to perform research and metadata construction. These instructions facilitate the intended functionality without attempting to bypass safety filters, escalate privileges, or hide malicious intent.

aso-research