The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to use an execute_code tool to modify operator properties such as nodeX, nodeY, nodeWidth, and nodeHeight. The documentation notes this is necessary because these attributes are properties of the operator object rather than parameters.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from the local environment via list_operators and get_connections to determine layout. 1. Ingestion points: list_operators and get_connections in SKILL.md read operator names and metadata. 2. Boundary markers: None identified. 3. Capability inventory: execute_code, reposition_operators, edit_operator, annotation, and set_parameters in SKILL.md allow for network modification. 4. Sanitization: None identified.

td-network-cleanup