The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external, potentially untrusted documents such as RFPs and company records, which could contain hidden instructions aimed at influencing agent behavior.
Ingestion points: Phase 1 extracts content from user-supplied RFP documents, company info, and track record files.
Boundary markers: The skill does not define specific delimiters or instructions to ignore malicious content embedded within the ingested documents.
Capability inventory: The skill performs file system operations, creating a _workspace/ directory and writing multiple analysis and proposal files (from 00_input.md to 06_review_report.md) across five different agents.
Sanitization: No explicit sanitization, validation, or filtering of the external document content is specified.

rfp-responder