skill-mcp-integration-manager
Warn
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill describes functionality for dynamic loading, runtime loading, and hot swapping of execution logic, which involves the manipulation and execution of code at runtime.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input data via the --input parameter in the execute-skill command, providing a vector for data-embedded instructions to influence agent behavior.
- Ingestion points: Data ingested via scripts/llmops_manager.py execute-skill.
- Boundary markers: None identified; no delimiters are used to separate user data from instructions.
- Capability inventory: Orchestration of external tools, sandbox management, and execution of registered skills.
- Sanitization: There is no documented evidence of input validation or sanitization protocols.
- [COMMAND_EXECUTION]: The skill requires the agent to execute a management script (scripts/llmops_manager.py) to facilitate core lifecycle operations such as skill registration and execution.
Audit Metadata