| Version | Supported |
|---|---|
| 2026.x.x | ✅ |
| < 2026.01.01 | ❌ |
We follow Calendar Versioning (YYYY.MM.DD). Security updates are provided for the latest release only.
If you discover a security vulnerability in the Research Workflow Assistant, please report it responsibly.
Use GitHub's private vulnerability reporting channel:
Security Advisories: https://github.com/andre-inter-collab-llc/research-workflow-assistant/security/advisories/new
Subject line: [SECURITY] RWA - Brief description
- Description of the vulnerability
- Steps to reproduce
- Potential impact
- Suggested fix (if you have one)
- Acknowledgment within 48 hours
- Assessment within 7 days
- Fix or mitigation timeline communicated after assessment
- Credit in the changelog and release notes (unless you prefer anonymity)
- Open a public GitHub issue for security vulnerabilities
- Exploit the vulnerability beyond what is necessary to demonstrate it
- Share the vulnerability with others before it has been resolved
The following areas are in scope for security reports:
- Exposure of API keys (NCBI, Zotero, OpenAlex, Semantic Scholar, CrossRef) through logs, error messages, or misconfiguration
- Insecure storage or transmission of credentials in
.envfiles or MCP server communications
- Injection vulnerabilities in MCP server tool inputs (search queries, DOIs, file paths)
- Path traversal in file-handling tools (Zotero local server, bibliography manager, PRISMA tracker, project tracker)
- Unauthorized file system access beyond intended project directories
- Unauthorized access to Zotero library data
- Leakage of Zotero API keys or user IDs
- Unintended exposure of local Zotero database or PDF contents
- Manipulation of PRISMA tracking data
- Tampering with project tracking records or decision logs
- Unauthorized modification of research data or citations
- Vulnerabilities in upstream dependencies (report these to the dependency maintainers)
- Vulnerabilities in the Zotero desktop application itself
- Issues requiring physical access to the user's machine
- Social engineering attacks
- Denial of service against external APIs (PubMed, OpenAlex, etc.)
- Never commit
.envfiles — they are in.gitignoreby default - Use the project virtual environment — never install into the system Python
- Keep API keys scoped — use read-only keys where possible (e.g., Zotero read-only access)
- Review AI-generated code before executing, especially analysis scripts
- Keep dependencies updated — run
pip install --upgradeperiodically within the venv