Repository navigation
RWA 2026.03.29
·
17 commits
to master
since this release
Script-first literature search workflow, Zotero attachment uploads, README Philosophy section, and infrastructure improvements.
Added
- Script-first search workflow:
draft_*_searchandrun_search_scripttools added to all five database MCP servers (PubMed, OpenAlex, Semantic Scholar, Europe PMC, CrossRef). Formal literature searches now generate a standalone Python script for user review before execution, improving reproducibility and auditability. - Shared search script engine:
generate_search_script,execute_search_script, andgenerate_and_run_scriptfunctions added torwa_result_storeshared library, powering the draft/run workflow across all servers. - Zotero attachment upload:
upload_attachmenttool added to zotero-server for uploading PDFs and other files to Zotero items via the Web API. - PDF upload utility scripts:
scripts/_upload_pdfs.pyandscripts/_upload_pdfs2.pyfor batch-uploading PDFs to Zotero library items. - Search results backfill utility:
scripts/backfill_search_exports.pyfor regenerating Excel exports from existingsearch_results.dbfiles. - Bug report issue template:
.github/ISSUE_TEMPLATE/bug-report.ymlfor structured issue reporting. - README Philosophy section: Added project philosophy explaining RWA as a proof of concept, the case against platform lock-in, and the existing research tech stack, with a link to the LinkedIn launch post.
- Search strategy template update: Added script-first workflow instructions to
templates/systematic-review/search-strategy.qmd. - Agent instructions for script-first workflow: Updated
systematic-reviewerandresearch-orchestratoragents with the mandatory draft-then-approve search protocol, including allowed direct MCP tool usage and DB-as-source-of-truth rules. - Global copilot instructions: Added Literature Search Protocol section enforcing the script-first, DB-as-source-of-truth workflow across all agents.
Changed
- MCP server project path handling: Improved error messages and validation for
project_pathparameters across PubMed, OpenAlex, Semantic Scholar, Europe PMC, and CrossRef servers. - API documentation: Updated
docs/api-setup-guide.md,docs/database-access.md, anddocs/getting-started.mdwith clarified instructions. - Search runners: Improved
search_runners.pyin the shared library for better script execution handling.
Fixed
- Chat parser: Handle non-string content in thinking blocks to prevent parsing errors in chat export.
Removed
- Obsolete
search_results.dbfiles from sample project directories (data now lives in project-specific paths). - Stale
prisma-flow.jsonfrom sample project (replaced by PRISMA tracker server workflow).