The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides the agent with several local Python scripts (research_engine.py, validate_report.py, verify_citations.py, md_to_html.py) to automate various stages of the research and report generation process. These scripts are executed via standard command-line interfaces.
[EXTERNAL_DOWNLOADS]: The verify_citations.py script performs network requests (HEAD and GET) to the doi.org metadata resolver and various external websites to verify citation accessibility and metadata (such as titles and publication years). These operations are essential for the skill's function and are performed with reasonable timeouts and user-agent headers.
[PROMPT_INJECTION]: The SKILL.md file contains complex instructional logic and protocols (e.g., Anti-Hallucination and Anti-Fatigue protocols) to ensure high-quality, grounded research. No instructions were found that attempt to override base model safety guidelines or exfiltrate sensitive data.
[DATA_EXFILTRATION]: Research reports and temporary state files are stored locally in the user's Documents and .claude directories. The skill does not demonstrate any patterns associated with reading or transmitting sensitive configuration files or credentials.

deep-research