Deep Research Pro v2

This skill appears to implement a reasonable multi‑phase research workflow and includes a small Python scoring script and report templates, but you should do the following before installing or relying on it: - Verify scoring consistency: the included QUALITY_CRITERIA and templates use a 0–1.0 or normalized quality scale, while scripts/quality-score.py computes totals on a 0–10 scale. Confirm how scores are normalized in the final report to avoid misleading grades. - Ensure runtime environment: the script expects Python3 and that the agent environment can read/write the described report directory structure and JSON source files. Test the script locally with sample source JSON to verify behavior. - Review web access implications: the skill will fetch many external sources (OpenAlex, PubMed, Google Scholar, company sites). Make sure the agent is allowed to make outbound web requests, and be aware of rate limits, robots.txt and copyright considerations (Google Scholar may require manual scraping or special handling). - Validate data handling: because the skill aggregates many sources and generates reports, check that no sensitive/internal documents will be pulled or published inadvertently when run in an environment with access to private intranets or proprietary APIs. - Test with non‑sensitive topics first: run the pipeline on public topics to confirm outputs (source lists, filtered/ excluded reasoning, final report) and that references/links are correctly formatted and accessible. Overall: coherent and low technical risk, but the scoring/template mismatch and heavy sourcing requirements merit review/testing before trusting outputs for decision making.