Skip to content

Security: ObtuseAI/dummy

SECURITY.md

Security policy

Dummy is a public-source, evidence-gated prediction-market intelligence and trading-research system. Its live execution paths are fail-closed, evidence-gated, and subject to explicit operator authorization (see the README safety invariants). Public visibility does not grant use, distribution, credential, broker, capital, or execution authority.

Supported versions

Security fixes are applied to the current 1.0.x release line. Historical snapshots are retained for auditability and are not separately supported.

Reporting a vulnerability

Report suspected vulnerabilities privately to the organization owner at cjharriskc@gmail.com. Do not open a public issue or pull request, and do not use any other public channel to describe a suspected or confirmed vulnerability.

Please include enough detail to reproduce the issue and, where relevant, the affected component. You will receive an acknowledgement and private coordination on remediation and disclosure timing.

Disclosure

Do not publish a suspected or confirmed vulnerability before coordinated review. Remediation and disclosure timing are handled privately with the organization owner; a public GitHub advisory may follow after a fix is available.

Safety posture

Consistent with the README, live execution remains fail-closed, evidence-gated, and operator-authorized. A live canary refuses to start until its evidence gate is satisfied, live orders pass through a hardened firewall adapter with transport-witnessed truth, and the kill file stops everything instantly and unconditionally. These controls are part of the security posture and must not be weakened to work around a report.

There aren't any published security advisories