The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources including opportunity rows and supporting conversations. These inputs could contain malicious instructions designed to influence the agent's analysis or its workspace mutation actions.\n- Ingestion points: references/task-method.md (opportunity records, firmographics, use cases, and supporting conversations).\n- Boundary markers: No explicit delimiters are mentioned to isolate untrusted data or warn the agent about potential instructions within the data.\n- Capability inventory: The skill possesses the ability to read and write to the workspace (via workspace-mutations.md) and utilize configured platform tools for analysis.\n- Sanitization: No explicit sanitization or filtering of the content within processed conversations is described in the task method or templates.

icp-refine