Skip to content

Security: Cerid-AI/homebrew-tap

Security

SECURITY.md

Security Policy

Supported Versions

Repository Supported
cerid-ai (latest main)
cerid-trading-agent (latest main)
cerid-boardroom (latest main)

Older pinned versions are not actively patched. Always run from main.


Reporting a Vulnerability

Do not open a public GitHub issue for security vulnerabilities.

Report security issues privately:

  1. Go to the affected repository on GitHub.
  2. Click SecurityReport a vulnerability (GitHub private advisory).
  3. Include: affected component, reproduction steps, potential impact, and any suggested fix.

We will acknowledge receipt within 48 hours and aim to provide a fix or mitigation within 14 days depending on severity.


Scope

In scope

  • Authentication bypasses or privilege escalation in the MCP server or API
  • Remote code execution via file ingestion or agent workflows
  • Data exfiltration from the knowledge base or secrets store
  • Dependency vulnerabilities in direct dependencies

Out of scope

  • Issues in third-party services (OpenRouter, Hyperliquid, Polymarket)
  • Vulnerabilities requiring physical access to the host machine
  • Social engineering attacks

Secrets & Encryption

Cerid AI uses age encryption for secrets at rest (.env.age). Age keys are stored outside the repo at ~/.config/cerid/age-key.txt. If you discover a committed secret or key material, report it immediately via the private advisory process above.


Dependency Security

Dependabot is enabled across all repos with weekly grouped updates. Lock files with hashes (requirements.lock, package-lock.json) are used to ensure reproducible builds. CodeQL analysis runs on every push.

There aren't any published security advisories