workflow-layout

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external workflow data via the panel_query_graph tool. This data (such as node titles, colors, or widget values) could theoretically contain malicious instructions designed to influence the agent's behavior.\n
  • Ingestion points: panel_query_graph (in SKILL.md end-to-end workflow section).\n
  • Boundary markers: None specified to distinguish data from instructions.\n
  • Capability inventory: panel_edit_node, panel_create_group, panel_move_group, panel_edit_group, panel_remove_group, panel_create_subgraph, panel_enter_subgraph, panel_exit_subgraph, panel_get_subgraph, panel_promote_widget, panel_move_rail, panel_expose_subgraph_output, panel_expose_subgraph_input, panel_unpack_subgraph, panel_save_workflow (in SKILL.md primitives section).\n
  • Sanitization: No explicit sanitization or validation of input data is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 01:58 PM
Security Audit — agent-trust-hub — workflow-layout