Skip to content

Security: yjsf216/feedback-agent

SECURITY.md

Security policy

Supported versions

Feedback Agent is currently pre-1.0. Security fixes are applied to the main branch and, after releases begin, to the latest release only. Older snapshots may not receive patches.

Report a vulnerability

Use GitHub private vulnerability reporting. Do not open a public issue for authentication bypasses, cross-app data access, credential exposure, remote code execution, injection, or vulnerabilities involving real user data.

Include the affected component and version or commit, reproduction steps, impact, and a minimal proof of concept. Remove API keys, production URLs, personal information, conversation content, and customer data before sending.

We aim to acknowledge a report within seven days and will coordinate a fix and disclosure when the issue is confirmed. These are best-effort targets, not a service-level agreement.

Security boundaries

  • Every domain operation must use a server-validated app ID.
  • App secrets, signing credentials, and model keys belong on trusted servers.
  • Production deployments must use HTTPS and exact CORS origins.
  • PostgreSQL, Redis, MinIO, and the MinIO console must not be exposed directly to the public internet.
  • AI-generated requirement drafts require human approval.

See docs/deployment.md for production controls.

There aren't any published security advisories