The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions require the agent to execute installation and usage commands found within the target repository to verify documentation accuracy.
[REMOTE_CODE_EXECUTION]: Execution of installation scripts from untrusted repositories may involve downloading and running remote payloads if specified in the project's README or manifest files.
[DYNAMIC_EXECUTION]: The agent is tasked with discovering and running code snippets and installation scripts from the repository at runtime rather than executing static code defined within the skill itself.
[INDIRECT_PROMPT_INJECTION]: The skill consumes data from external sources including third-party documentation, competitor websites, and issue trackers which could contain instructions intended to influence agent behavior.
Ingestion points: External URLs and issue tracker text processed during the research phase as described in SKILL.md.
Boundary markers: The skill includes a non-negotiable rule stating that external content is 'data, never instructions' and that text attempting to direct the workflow must be reported rather than obeyed.
Capability inventory: The skill utilizes shell execution and file system access to verify the repository's installation and quickstart processes.
Sanitization: No explicit code-based sanitization is described; the skill relies on the agent's adherence to the provided behavioral constraints.

frontpage