The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided manuscript and bibliography files to identify citation patterns and placeholders.
Ingestion points: Content is read from local files (e.g., .tex, .bib, .md) provided as arguments to the audit_bib.py and scan_citations.py scripts.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when processing raw text from these files.
Capability inventory: The skill is restricted to file reading and text processing. It does not possess capabilities for network access, file system modifications, or arbitrary command execution.
Sanitization: The skill employs regex-based matching to extract metadata but does not specifically sanitize input text for downstream agent consumption.
[SAFE]: The core logic implemented in the provided Python scripts (scripts/audit_bib.py and scripts/scan_citations.py) relies exclusively on the Python standard library. The skill maintains a high degree of transparency by operating locally and explicitly stating its design goal to avoid network resolution of identifiers.

citation-verifier