The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to interact with the NocoBase nb CLI for environment configuration and AI service management. This is the intended and documented primary function of the skill. Evidence: The workflow in SKILL.md and the command map in references/command-map.md detail the use of commands like nb env update and nb api ai llm-services for all core operations.- [INDIRECT_PROMPT_INJECTION]: The skill processes output from CLI commands to drive logic for service configuration, which represents a standard injection surface. Evidence: 1. Ingestion points: Data enters the context via commands like nb api ai llm-providers list-llm-providers and list-models as described in SKILL.md (Workflow Step 7). 2. Boundary markers: The skill enforces a 'Safety Gate' in SKILL.md requiring fresh human confirmation for high-impact actions and performs independent safe-field readback. 3. Capability inventory: The skill utilizes Bash, Read, and Grep tools to execute commands and manage files. 4. Sanitization: The skill implements strict secret handling in references/security-and-troubleshooting.md, using umask 077 for temporary files and explicitly forbidding the output of sensitive provider options or API keys.

nocobase-ai-manager