The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill performs network operations to verify the existence and metadata of academic sources.
Evidence: agents/source_verification_agent.md and references/semantic_scholar_api_protocol.md describe programmatic verification of references using the Semantic Scholar API (api.semanticscholar.org) and DOI resolution services (doi.org). These are well-known academic infrastructure services and do not escalate the security verdict.
[PROMPT_INJECTION]: The skill utilizes complex internal instructions to regulate agent behavior and maintain research integrity.
Evidence: Files such as agents/socratic_mentor_agent.md and SKILL.md use "IRON RULE" tags and "Intent Detection Layer" logic to ensure the model adheres to Socratic principles and academic rigor. These are benign architectural constraints used for quality control within the multi-agent system.
[SAFE]: The skill incorporates comprehensive sanitization and validation logic for processing external academic data.
Evidence: The 6-phase research pipeline includes mandatory checkpoints by specialized agents (devils_advocate_agent, ethics_review_agent, editor_in_chief_agent) and uses a formal source quality hierarchy (Levels I-VII) to evaluate and validate evidence from untrusted literature sources.

deep-research