discord-hardened
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes comprehensive defensive instructions that explicitly forbid the agent from acting on fabricated or injected context within Discord messages. This serves as a strong mitigation against indirect prompt injection by requiring direct operator confirmation for actions initiated by message content.
- [COMMAND_EXECUTION]: The tool provides extensive control over Discord servers, including moderation and channel management. These capabilities are restricted by mandatory safety guardrails requiring explicit operator verification for destructive actions like message or channel deletion.
- [DATA_EXFILTRATION]: While the skill facilitates reading message histories and server metadata, the instructions emphasize human oversight and prevent the autonomous collection or redistribution of data.
- [EXTERNAL_DOWNLOADS]: The skill supports uploading media from remote URLs and local file paths (e.g., file:///tmp/). Potential risks associated with local file access or remote fetching are mitigated by instructions requiring the operator to review each item individually before processing.
Audit Metadata