research-mental-check

Pass

Audited by Gen Agent Trust Hub on Apr 25, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by instructing the agent to read historical log files from '~/phd-log/mental-checkins/' to identify recurring patterns. This allows potentially untrusted content from past interactions to influence current agent logic.
  • Ingestion points: Historical markdown files located in the user's home directory.
  • Boundary markers: Absent; the skill does not define specific delimiters to isolate historical data from active prompts.
  • Capability inventory: The skill utilizes file system read and write operations.
  • Sanitization: Absent; log data is processed without escaping or validation.
  • [COMMAND_EXECUTION]: The skill performs local file system operations to manage personal logs in the '~/phd-log/' directory. This involves the storage of sensitive personal reflections in plain-text markdown files on the user's local disk.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 25, 2026, 11:59 AM