The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install an external Python package named gateguard-ai from a package registry (PyPI) to access full configuration features.
[COMMAND_EXECUTION]: The documentation references the use of a CLI tool gateguard via gateguard init for project-level setup and configuration.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection (Category 8) because it mandates the agent read and analyze codebase files and data structures before proceeding with edits.
Ingestion points: The agent is instructed to use Grep, Glob, and file reading tools to extract facts from the local repository (e.g., imports, public functions, and data schemas in files).
Boundary markers: The instructions do not define specific delimiters or encapsulation techniques for the gathered facts to distinguish them from the agent's instructions.
Capability inventory: The skill is designed to manage and gate the agent's use of Edit, Write, and Bash tools.
Sanitization: The skill includes a policy to use "redacted or synthetic values" for data schemas to prevent data exposure, but it lacks specific sanitization or filtering to prevent malicious instructions embedded in those files from influencing the agent's behavior.
[PROMPT_INJECTION]: The skill provides an environment variable override (ECC_GATEGUARD=off) to bypass its own gating logic, which could be exploited to disable the safety features of the skill if the agent is instructed to do so.

gateguard