Skip to content

Security: orchestratemcp/OrchestrateKIT-MCP

SECURITY.md

Security policy

Supported versions

Until the first public tag is published, security fixes target the current master branch. After v0.1.0, the latest tagged release and master are the supported lines unless a release note says otherwise.

Reporting a vulnerability

Please use GitHub's Report a vulnerability flow in the repository Security tab. It opens a private advisory visible only to the maintainers.

Do not open a public issue for a suspected vulnerability, exposed credential, or report containing private workflow data. If private reporting is temporarily unavailable, open a content-free issue asking a maintainer to enable a private contact channel.

Include only what is needed to reproduce and assess the problem:

  • affected commit or version;
  • impact and prerequisites;
  • minimal reproduction steps;
  • suggested mitigation, if known.

Never include real API keys, OAuth tokens, customer data, prompts, generated plans, or private registry/LAB data. Use synthetic examples and revoke any credential that may have been exposed.

Security boundary

OrchestrateMCP is a read-only, stateless planning service. It does not store workflow credentials, run planned workflows, or make LLM calls inside MCP tools. Generated workflow examples may require third-party credentials; those credentials belong in the target platform's secret manager and must never be committed to this repository.

This boundary reduces risk but is not a guarantee that the software is free of vulnerabilities. Reports involving MCP transport handling, unsafe generated instructions, secret exposure, authorization assumptions, or dependency vulnerabilities are in scope.

There aren't any published security advisories