The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted document data, which presents a surface for indirect prompt injection attacks where malicious content within a document could attempt to influence agent behavior.
Ingestion points: External files are ingested through the document_paths parameter in the provenance_analysis, anomaly_detection, and redaction_detection capabilities in SKILL.md and schema.yaml.
Boundary markers: The skill does not define specific delimiters or instructions to treat ingested document text as untrusted data within the prompt context.
Capability inventory: The skill utilizes Bash, Read, Glob, and Grep tools, providing a wide functional surface for potential exploitation.
Sanitization: The instructions do not specify any sanitization, escaping, or validation of the text content extracted from documents.
[EXTERNAL_DOWNLOADS]: The skill requires several standard Python libraries and system tools for its stated forensic analysis functionality.
Evidence: The configuration in schema.yaml specifies dependencies on pdfplumber, PIL (Pillow), cv2 (OpenCV), dateutil, and the system utility exiftool. These are well-known and standard tools for document and image processing.

document-forensics