The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external text data through the xberg CLI and library, which is a standard characteristic of document processing utilities.
Ingestion points: Text is ingested via the xberg extract command and the --text argument in xberg chunk as seen in SKILL.md.
Boundary markers: The documentation does not prescribe specific delimiters or instruction-isolation markers for the processed content.
Capability inventory: The skill involves shell command execution (xberg) and local file system access (cat, extract).
Sanitization: No explicit sanitization or filtering logic is provided in the documentation examples.
[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill mentions the use of HuggingFace tokenizer IDs (e.g., Xenova/gpt-4o). These resources are fetched from a well-known service provider to configure the splitting logic of the tool.
[DATA_EXPOSURE_AND_EXFILTRATION]: The examples demonstrate reading local files such as PDF and Markdown reports. These operations are local to the user's environment and align with the tool's primary purpose of text extraction and chunking.

chunking