The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides and documents the use of subprocess_async_client and subprocess_sync_client, which execute the litestar CLI in a background process for end-to-end testing.
Evidence: Found in SKILL.md and references/async_testing.md within the "Live Subprocess Clients" and "Live Subprocess Server" sections.
[INDIRECT_PROMPT_INJECTION]: The skill defines a vulnerability surface by demonstrating how to write code to the file system and subsequently execute it using a subprocess helper, which could be exploited to run arbitrary code if an agent is manipulated by untrusted data.
Ingestion points: User-provided handler logic and application configurations in SKILL.md and references/async_testing.md.
Boundary markers: Absent in the provided examples.
Capability inventory: Includes file system writing (Path.write_text in examples), subprocess execution (subprocess_async_client), and internal network requests (AsyncTestClient).
Sanitization: No explicit sanitization or validation of the generated code strings before execution is demonstrated.

litestar-testing