The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the subprocess.run method in formats.py to invoke the ebook-convert command-line utility (part of the Calibre suite). This is used to handle conversion for .mobi and .azw3 formats. The implementation uses list-based arguments without a shell, which follows security best practices for executing external commands.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes and extracts text from external ebook files and highlight logs.
Ingestion points: Untrusted data is ingested from user-provided Kindle ebook files (.epub, .mobi, .azw3) and the My Clippings.txt file via clippings.py and formats.py.
Boundary markers: The skill appends an  marker to the extracted text to signal completion to downstream tools, but it does not include explicit delimiters or instructions to ignore potential commands embedded within the book text.
Capability inventory: The skill has the capability to execute subprocesses (ebook-convert) and write extracted text and metadata to the local filesystem in the ~/clawd/library/books/ directory.
Sanitization: The skill performs basic HTML tag stripping and sanitizes filenames using a slugification function, but it does not filter or sanitize the extracted natural language content for malicious instructions.

ingest-kindle