The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute the geo CLI. The geo score <url> command interpolates a user-provided URL directly into a shell command. While typical for agent tools, this creates a surface for command injection if the platform's shell execution environment does not adequately sanitize arguments.
[EXTERNAL_DOWNLOADS]: The skill requires the seomator-geo-audit package to be installed on the host system. The instructions guide users to install this from external registries (PyPI via uv or pipx). This introduces a dependency on external code not contained within the skill itself.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from external websites which could contain malicious instructions designed to subvert the agent's behavior.
Ingestion points: Data is ingested through the output of geo score <url> and geo fetch <url> (referenced in SKILL.md and sections/remediation.md).
Boundary markers: Present. The skill contains explicit instructions to treat excerpts from findings[].excerpt and signals[].detail as untrusted data and to report any embedded instructions as findings rather than following them.
Capability inventory: The agent has Bash shell access, which could be exploited if an indirect injection successfully bypasses the instructions.
Sanitization: The skill relies on the CLI tool's normalization but does not perform additional sanitization of the external text before processing.

citability