pbc-package

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill requires extracting and outputting the "first meaningful text snippet" and other file-derived summaries/responses (including filenames/paths), which can cause any secrets contained in input files to be emitted verbatim by the LLM.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In the required workflow, the outsider-authored “PBC request list” (XLSX/CSV/PDF) is parsed and its free-text description is directly read by scripts/match_pbc.py for fuzzy scoring against filenames/summaries, so an outsider can submit poison via the monitored input artifact without selecting a specific item.