data-ingest

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to transcribe source content verbatim (including images) into wiki pages and to read .env, so any API keys, tokens, or passwords present in the ingested files would be included verbatim in generated output.

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

• Third-party content exposure detected (high risk: 0.75). The skill ingests “arbitrary text data” from user-specified source files (e.g., chat exports/logs/transcripts/HTML/markdown/plaintext) via the Read tool, and that content is explicitly treated as untrusted free text to be distilled into wiki pages—so if the source is outsider-authored, its prose will be fed into the agent’s LLM context.