personal-assistant

Pass

Audited by Gen Agent Trust Hub on Feb 23, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): No patterns associated with prompt injection, safety bypasses, or instructions to ignore system rules were found in the provided instructions.
  • Data Exposure & Exfiltration (SAFE): The skill's documentation indicates that data is stored locally at ~/.claude/personal_assistant/. The code does not contain any network operations (like curl or requests) to transmit data externally.
  • Obfuscation (SAFE): All provided files are in plain text. No encoded strings, homoglyphs, or hidden characters were detected.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill uses no external Node.js or Python packages. It relies on standard library modules (e.g., json, datetime, pathlib) and local script references.
  • Privilege Escalation (SAFE): No commands requiring elevated privileges (sudo) or modifications to system files were identified.
  • Persistence Mechanisms (SAFE): The skill does not attempt to create cron jobs, startup entries, or modify shell profiles.
  • Indirect Prompt Injection (SAFE): While the skill processes user-inputted task descriptions, this functionality is standard for its purpose and does not include risky capability pairings like arbitrary command execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 23, 2026, 05:28 AM