The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/harvest_complaints.py utilizes subprocess.run to execute the system utilities pdfinfo and pdftotext. These commands are used to verify page counts and extract text content from downloaded PDF documents for further processing.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes large volumes of external data from legal vendors, which creates an attack surface for instructions potentially embedded in document titles or docket entries. 1. Ingestion points: API responses from Lex Machina and Docket Alarm processed in scripts/harvest_complaints.py and scripts/da_grab.py. 2. Boundary markers: No specific delimitation or 'ignore instructions' markers are used when processing external document text. 3. Capability inventory: The skill has the capability to perform network requests via requests and execute local system commands via subprocess.run. 4. Sanitization: External document titles are normalized for whitespace and case in scripts/harvest_complaints.py using the norm() function, but do not undergo extensive safety filtering or escaping.

court-dockets