The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/line_insert.py via the command line to perform precise code insertions at specific line numbers.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to move code between files, which involves reading untrusted content and writing it to other locations on the filesystem.
Ingestion points: The code argument passed to line_insert.py and content handled by Read and Edit tools.
Boundary markers: None implemented; the skill relies on the agent's context for delimitation.
Capability inventory: The skill has the ability to read, write, and grep files, as well as execute local scripts.
Sanitization: The line_insert.py script includes validation to prevent directory traversal by checking for .. in file paths and resolving absolute paths.
[DYNAMIC_EXECUTION]: The skill documentation describes an "Execution Mode" for batch operations involving the execution of Python code that interacts with internal agent APIs (api.filesystem, api.code_analysis).

code-transfer