The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect instructions as it processes untrusted conversation context to generate documents. Ingestion points: Processes current conversation context and explores the local repository (SKILL.md). Boundary markers: The instructions lack explicit delimiters or warnings to ignore instructions embedded within the conversation history. Capability inventory: The agent is instructed to use external tools to publish content to a project issue tracker. Sanitization: No specific sanitization or filtering logic is defined for the synthesized output.

to-spec