The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to ingest and analyze untrusted external story content, which is a standard surface for indirect prompt injection. While the current capabilities are limited, embedded instructions in the story could attempt to bias or subvert the analysis process. Ingestion points: Story content enters the workflow via the 'Read' tool in the 'Text Preprocessing' and 'Parallel Analysis' phases. Boundary markers: The workflow does not define specific delimiters or instructions to the agent to ignore commands within the source text. Capability inventory: The skill only specifies the 'Read' tool, minimizing potential side effects. Sanitization: No sanitization or input filtering is specified for the dramatic analysis process.

drama-workflow