The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @llamaindex/liteparse package from the official npm registry using npx or pnpm. This is a well-known utility from the LlamaIndex organization.
[COMMAND_EXECUTION]: The skill executes local shell commands via the lit CLI to parse documents, generate screenshots, and perform OCR operations. These operations are scoped to the files provided by the user.
[INDIRECT_PROMPT_INJECTION]: As the skill ingests and processes untrusted external documents (PDFs, Word files, etc.) to extract text, it provides a surface for indirect prompt injection. If the extracted content is subsequently processed by an AI agent without sanitization, it could influence the agent's behavior.
Ingestion points: lit parse and lit batch-parse commands process various unstructured file formats from the local file system.
Boundary markers: None specified in the instructions; the skill returns raw extraction results.
Capability inventory: The skill has the capability to write files to disk (-o flag) and execute local commands for parsing and OCR.
Sanitization: The skill relies on the underlying LiteParse library for extraction; no additional sanitization of the extracted text is described.

liteparse