session-pretty-replay

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • Prompt Injection (LOW): The skill is vulnerable to indirect prompt injection via the consumption of untrusted session data.\n
  • Ingestion points: Processes external JSON/JSONL files (e.g., chat.jsonl) containing conversation history.\n
  • Boundary markers: Employs visual box separators but lacks semantic boundary instructions to ensure the agent treats the content as inert data and ignores embedded instructions.\n
  • Capability inventory: Accesses the file system for reading logs and writing formatted output.\n
  • Sanitization: No content filtering or sanitization of the processed log data is implemented in the instruction set.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:21 PM