Query Doctor handles operational diagnostics and may run near sensitive query, profile, metadata, and cluster-management systems. Treat all raw collected data as sensitive unless it has passed the project redaction and trust boundaries.

Until a formal release-support policy is added, security-sensitive reports should be handled privately with the repository maintainers.

Use GitHub's private "Report a vulnerability" flow for exploitable vulnerabilities or trust-boundary failures. Share only sanitized reproduction details until a private channel is confirmed.

Do not file public issues containing:

• raw SQL or query text;
• raw Impala profiles;
• raw provider API JSON or daemon responses;
• raw metadata output;
• hostnames, IP addresses, usernames, emails, principals, tokens, cookies, passwords, Authorization headers, embedded URL credentials, local config contents, or production profile text;
• local paths or generated artifact contents from real environments.

When reporting a security issue, include a minimal sanitized reproduction, affected command or workflow, expected behavior, actual behavior, and why the output crosses a trust boundary.

For public issues, describe the affected boundary and use synthetic placeholders only. Do not attach real generated artifacts from production environments.

Important project invariants:

• Python/analyzer facts are the only trusted diagnostic evidence.
• Raw LLM output is untrusted unless validated.
• Browser-visible UI and trusted reports must not expose raw operational data.
• External collection must be explicit, bounded, read-only, redacted, and safe by default.
• Query Optimizer must not execute pasted SQL.
• Metadata collection must stay read-only and allowlisted.

See docs/safety-contract.md for the complete contract.