reflection-journal
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts unstructured brain dumps from the user and stores them for later theme extraction and pattern recognition. This workflow creates an attack surface where instructions embedded within a journal entry could potentially be interpreted as instructions by the agent during subsequent review or analysis sessions. \n
- Ingestion points: Raw user brain dumps as described in the 'Input: Accept Dumps' section of SKILL.md. \n
- Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from following directives contained within the journaled content. \n
- Capability inventory: The skill performs file system write operations to the
journal/directory and reads these files to perform 'Pattern Recognition Across Entries'. \n - Sanitization: There is no mention of sanitizing or escaping user-provided text before it is stored or analyzed.
Audit Metadata