The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (HIGH): The skill's installation instructions in SKILL.md explicitly command the agent to download a script and pipe it directly into the bash shell. Since the source organization 'metalbear-co' is not in the trusted list, this is a high-severity risk for arbitrary code execution.
Evidence: curl -fsSL https://raw.githubusercontent.com/metalbear-co/mirrord/main/scripts/install.sh | bash
[EXTERNAL_DOWNLOADS] (MEDIUM): The skill attempts to download and install packages from untrusted external sources (Homebrew tap and GitHub).
Evidence: brew install metalbear-co/mirrord/mirrord and the curl-based install script.
[COMMAND_EXECUTION] (LOW): The skill invokes a local command-line interface (mirrord) to perform validation on configuration files.
Evidence: mirrord verify-config /path/to/config.json in SKILL.md.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection because it ingests untrusted JSON configurations from users and processes them through an external CLI tool without explicit sanitization beyond schema validation.
Ingestion point: User-provided mirrord.json content.
Boundary markers: Not explicitly defined for the input data.
Capability inventory: Shell execution via mirrord verify-config.
Sanitization: Mentions strict JSON parsing but relies on the external CLI for final validation.

mirrord-config