The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to use shell commands, specifically uv run and the wedge CLI, to build, lock, and verify Python-based helper tools within a repository.
[DYNAMIC_EXECUTION]: The skill facilitates the creation of executable Python ZIP archives (.pyz) from local source templates. These archives are then loaded and executed by a generated launcher to perform data processing tasks.
[INDIRECT_PROMPT_INJECTION]: The skill defines workflows for reading and processing external data, which introduces a potential attack surface.
Ingestion points: The provided Python template (assets/records.py) uses path.read_text() to ingest data from local files.
Boundary markers: The skill utilizes the fromargs library, which enforces output limits via decorators (e.g., @app.command(limit=2)) and ensures results are returned as structured JSON to prevent context poisoning.
Capability inventory: The skill has the ability to execute shell commands and launch generated Python archives through the uv and wedge tools.
Sanitization: Data processing logic in the template (line stripping and frequency counting) transforms raw input into structured counts, providing a layer of abstraction from the original text.

wedge