mega-code-login
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: Executes shell commands and Python modules via
uv runto perform OAuth steps and local setup.\n- [DATA_EXFILTRATION]: Connects tohttps://console.megacode.aito conduct session creation and polling for authentication completion. This is the intended behavior for the specified service.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by parsing JSON command output and using the values in subsequent bash commands. In accordance with the analyzer rules, the severity is reduced to SAFE as this interaction is required for the skill's primary purpose.\n - Ingestion points: JSON output containing
client_idandbase_urlfrom the initial setup step.\n - Boundary markers: None.\n
- Capability inventory: Shell command execution via
uv runandbash.\n - Sanitization: None performed on the parsed values before interpolation.
Audit Metadata