nocobase-notification-manage
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes data from external sources such as notification send logs (notificationSendLogs) and in-app message records (notificationInAppMessages), which may contain untrusted user-generated content.\n
- Ingestion points: The skill uses nb api resource list and get commands in references/diagnostics.md to fetch logs and message details.\n
- Capability inventory: The agent is granted Bash, Read, and Grep permissions, enabling it to perform system-level actions based on its instructions.\n
- Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the processed data.\n
- Sanitization: The instructions do not specify a requirement to sanitize or escape data fetched from the API before it is interpreted or displayed.\n- [CREDENTIALS_UNSAFE]: The skill facilitates the configuration of notification channels that require sensitive information, including SMTP passwords.\n
- Evidence: references/channel-configuration.md provides command examples for creating email channels using the nb CLI, which include a password field within the JSON payload.\n
- Mitigation: SKILL.md and references/channel-configuration.md explicitly mandate the redaction of passwords and secrets in all output provided to the user.\n- [COMMAND_EXECUTION]: The skill uses the Bash tool to interact with the NocoBase environment via the nb CLI.\n
- Evidence: Multiple commands such as nb api resource list, nb api resource create, and nb api resource update are used throughout the reference files.\n
- Context: These commands are scoped to the nb utility for managing platform resources and appear consistent with the skill's stated administrative purpose.
Audit Metadata