The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for fetching and installing external content from the GitHub repository registry using the platform's native /skill install command interface.
[COMMAND_EXECUTION]: The skill utilizes a set of privileged commands (/skill trust, /skill update, /skill uninstall) that modify the local filesystem at ~/.mimofan/skills and change the agent's execution environment.
[PROMPT_INJECTION]: The skill manages untrusted third-party code, which represents a potential surface for indirect prompt injection attacks where malicious instructions are embedded in the installed skill files.
Ingestion points: Community-contributed SKILL.md files and accompanying scripts downloaded from GitHub or provided locally.
Boundary markers: The instructions include explicit guardrails, such as "Treat scripts and companion files as untrusted until reviewed" and requiring manual inspection before trust is granted.
Capability inventory: The skill manages the lifecycle of other agent extensions, including installation, activation, and granting execution trust for scripts and tools.
Sanitization: Does not implement technical sanitization; instead, it relies on instructions for the agent to perform manual verification and requires explicit user consent for execution.

skill-installer