brain-dump
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is specifically designed to ingest 'messy stream-of-consciousness' data, transcribed voice memos, and long blocks of mixed text from the user. This creates a surface for indirect prompt injection if the user pastes content containing hidden instructions.
- Ingestion points: The skill reads input from the user (typed or transcribed) and also scans all files in the working folder, including project files, past brain dumps, and journals (
SKILL.md, Step 1). - Boundary markers: There are no specific instructions to use delimiters or to ignore embedded instructions within the processed text.
- Capability inventory: The skill has the capability to write files to the local filesystem (e.g.,
brain-dump-YYYY-MM-DD-HHMM.md) and is encouraged to suggest 'one click away' next steps which could involve further tool usage (SKILL.md, Step 3). - Sanitization: No evidence of sanitization or filtering of the input data is present.
Audit Metadata