The Agent Skills Directory

[COMMAND_EXECUTION]: The skill describes the use of the ts-pack CLI tool for source code analysis. It also includes examples utilizing common development utilities like jq for JSON manipulation and git ls-files for identifying project files.
[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to ingest external source code files. Comments and docstrings within these files could potentially harbor instructions intended to influence the agent's behavior during the analysis phase.
Ingestion points: The ts-pack process <file> command in SKILL.md allows the agent to read and process content from local source files.
Boundary markers: The instructions do not specify any delimiters or warnings to the agent regarding the untrusted nature of the content found within docstrings or comments.
Capability inventory: The skill uses the ts-pack CLI and the tree_sitter_language_pack Python library for metadata extraction.
Sanitization: While the output is structured as JSON, there is no explicit mention of sanitizing the raw string content (such as docstrings) extracted from the files.

extracting-code-structure