The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill directs the agent to parse and execute commands from repository files, which serve as untrusted data sources. * Ingestion points: Configuration files including Makefile, package.json, go.mod, pom.xml, pyproject.toml, and .github/workflows/*.yml (identified in SKILL.md). * Boundary markers: None. The instructions do not specify any delimiters or warnings to ignore instructions within these files. * Capability inventory: The skill grants the agent the ability to run shell commands found in the repository. * Sanitization: None. There are no requirements to validate or escape the command strings before they are passed to the shell.
[DYNAMIC_EXECUTION]: The skill facilitates the execution of commands that are dynamically determined based on the contents of the target repository, allowing for code execution that is not statically defined in the skill's instructions.

verify