lgnh-agent-md-updator
Fail
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to perform file system operations, including reading, writing, and creating backups, within sensitive configuration directories such as
~/.claude/,~/.agents/, and~/.gemini/. - [PROMPT_INJECTION]: The skill is designed to modify the persistent instruction files (system prompts) that define the agent's core behavior. This provides a mechanism for establishing persistence for malicious instructions or behavioral overrides that bypass safety protocols across all future interactions.
Recommendations
- AI detected serious security threats
Audit Metadata