session-learn

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted session data (user conversations) and persists extracted information into a wiki that the agent likely references in future sessions. This could allow a malicious user to inject persistent instructions or false information into the agent's long-term memory.
  • Ingestion points: The entire conversation history and tool execution logs are scanned for knowledge signals (SKILL.md).
  • Boundary markers: No explicit delimiters or instructions are used to distinguish untrusted user input from verified knowledge during the extraction process.
  • Capability inventory: The skill performs comprehensive file system operations, including creating and modifying files in the wiki/, decisions/, creativity/, and system/ directories.
  • Sanitization: The skill performs deduplication but does not mention sanitizing or validating the content extracted from the session.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 07:02 PM