The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/find_skills.py interacts with the system using subprocess.run to call python3 and npx. These commands are executed using argument lists rather than shell strings, preventing command injection vulnerabilities.
[REMOTE_CODE_EXECUTION]: The skill facilitates the installation of remote code by design. It uses the npx skills add command to fetch and install skills from a specified repository, defaulting to the author's official volcengine/volcengine-agent-toolkit repository.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a gateway for installing new capabilities based on user input, creating an attack surface for indirect prompt injection.
Ingestion points: The skill ingests untrusted data from user requests and structured data from the local references/catalog.json file.
Boundary markers: There are no explicit markers used to delimit untrusted data during the discovery process.
Capability inventory: The skill possesses the capability to modify the agent's environment by installing new scripts and tools via scripts/find_skills.py.
Sanitization: Input validation is performed by the resolve_skills function, which ensures only skills listed in the local catalog can be targeted for installation.

volcengine-find-skills