optimize-application

Pass

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection surface. The skill ingests untrusted data while possessing significant configuration management capabilities.\n
  • Ingestion points: The skill reads external data via branchly-app_read_session_detail (user chat history) and branchly-app_read_node (scraped website content).\n
  • Boundary markers: Absent. There are no instructions to use delimiters or "ignore embedded instructions" markers when the agent processes this external data.\n
  • Capability inventory: The skill can modify the application's core logic using branchly-app_update_data_source, branchly-app_update_tool, and branchly-app_create_prompt.\n
  • Sanitization: Absent. No sanitization or validation of the ingested data is performed before it is used to influence the agent's decision-making process for application updates.\n- [COMMAND_EXECUTION]: The skill makes extensive use of vendor-specific MCP tools (e.g., branchly-app_read_sessions, branchly-app_update_data_source) to interact with the application environment. These operations are intended for managing application state and configuration but involve reading and writing sensitive application data.
Audit Metadata
Risk Level
SAFE
Analyzed
May 7, 2026, 08:04 AM
Security Audit — agent-trust-hub — optimize-application