The Agent Skills Directory

[COMMAND_EXECUTION]: The script install_all_plugins.py orchestrates bulk installations by executing the bridge_installer.py script via subprocess.run. The command is constructed as a list, which is a standard safety measure to prevent shell injection vulnerabilities.
[PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by ingesting logic from external plugin files and incorporating them into agent system instruction files (e.g., CLAUDE.md, .github/copilot-instructions.md).
Ingestion points: Markdown files and TOML files located within the plugins/ directory.
Boundary markers: The installer wraps injected content in  markers for management and idempotency.
Capability inventory: The skill has access to file system read/write operations and the ability to execute sibling Python scripts.
Sanitization: Content is transformed for target environment compatibility (e.g., actor name swapping) but is not otherwise sanitized for malicious content prior to being written into agent configuration files.

bridge-plugin