The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied topics and ideas to generate content plans, creating a surface for indirect prompt injection attacks where malicious instructions could be embedded in the input.
Ingestion points: User input provided via the /series command (e.g., [topic or idea]) is interpolated into the workflow in SKILL.md during the ASSESS, DECIDE, and GENERATE phases.
Boundary markers: The instructions lack explicit delimiters or instructions for the agent to treat the user-provided topic as untrusted data or to disregard any embedded instructions within the topic string.
Capability inventory: The skill is configured with a broad set of capabilities including Bash, Write, Edit, Read, Grep, Glob, and Task tools.
Sanitization: No sanitization, validation, or escaping logic is defined to mitigate the risk of user-provided content influencing the agent's behavior or tool usage.

series-planner