| Version | Supported |
|---|---|
| 1.0.x | Yes |
| <1.0 | No |
Please report security-sensitive issues privately rather than opening a public issue. Use GitHub's private vulnerability reporting when available.
We aim to acknowledge reports within 48 hours and provide a resolution timeline within 7 days.
- Never include API keys, access tokens, or credentials in issues, logs, screenshots, fixtures, or example repositories
- citeguard must not print environment-variable values or request authorization headers
- Custom base URLs containing credentials are sanitized before printing in CLI output, logs, and reports
- Reports and cache files must contain only data required for citation analysis
- API keys are never written to generated reports or cache metadata
citeguard is a local CLI, but the analysis pipeline may send limited document-derived content to configured third-party APIs depending on which features are enabled:
- Academic metadata providers: search queries derived from citation text and bibliography entries; never full document prose
- LLM providers (when configured and enabled): paragraph text for claim extraction, entailment classification, and source matching
- Network usage is reported per-run via
ExecutionContextfields (academic_network_used,llm_network_used) so you can verify whether any remote calls were made - When running in offline/cache-only mode, no data leaves the local machine
Review the Privacy section in README.md before processing confidential or unpublished documents.
This security policy applies to the citeguard CLI tool and its official GitHub repository. It does not cover third-party academic metadata providers or LLM APIs used by the tool.