anki-connect

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): No instructions attempt to override agent constraints or extract system prompts. The instructions are strictly focused on API interaction and safety protocols.
  • Data Exposure & Exfiltration (SAFE): Network operations are restricted to the local loopback address (127.0.0.1). No sensitive local file paths are accessed, and no hardcoded credentials are present.
  • Obfuscation (SAFE): The file contains clear, readable markdown and standard bash commands. No Base64, zero-width characters, or encoding tricks were detected.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill relies on standard system utilities (curl and jq). It does not download external scripts or packages from the internet.
  • Privilege Escalation (SAFE): No commands require sudo or attempt to modify system-level permissions.
  • Persistence Mechanisms (SAFE): The skill does not attempt to create cron jobs, modify shell profiles, or establish startup persistence.
  • Indirect Prompt Injection (SAFE): While the skill reads data from a local database (Anki), it implements a strict 'Safety and Confirmation Policy' that requires human-in-the-loop verification for all state-changing operations, mitigating risks from potentially malicious note content.
  • Command Execution (SAFE): The skill uses structured JSON generation with jq to interact with a local API. This is the intended purpose of the skill and is implemented using best practices to avoid shell injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:56 PM