The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill processes external DAG data and node configurations (e.g., priority, deadlines) which are used to determine scheduling logic. If these inputs are sourced from untrusted or attacker-controlled files, they could potentially influence the agent's behavior.
Ingestion points: The skill uses Read and Glob tools to ingest DAG structure and node configurations (node.config).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands in the DAG data are present.
Capability inventory: The skill possesses Write and Edit permissions, allowing it to modify files based on its scheduling logic.
Sanitization: No evidence of sanitization or validation of the input data fields (like nodeId or priority) before use in logic or potential interpolation into other prompts.

dag-task-scheduler