You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
Project summary
AI agents increasingly take real-world actions — calling tools, moving data, executing workflows — faster than any human can supervise. Synchronicity is an external governance authority that sits outside the agent and evaluates each proposed action against explicit, versioned policy before it executes, returning a signed, logged decision: ALLOW, HOLD, or DENY.
It never executes actions itself (strict separation of governance from execution), it is default-deny and fail-closed (unknown or unverifiable requests are blocked, not waved through), and every decision produces cryptographically verifiable, replayable audit evidence. The design aligns with the NIST SP 800-207 Zero Trust Policy Decision Point model, applied to autonomous AI.
What are this project's goals? How will you achieve them?
The goal is to demonstrate that agent action-authorization can be made deterministic, external, and auditable — a missing control layer for safely deploying autonomous agents.
This funding supports two public artifacts:
A reference implementation demonstrating pre-execution ALLOW/HOLD/DENY decisioning, with signed, replayable audit records.
An open evaluation harness that lets others test the governance layer's determinism, fail-closed behavior, and auditability across representative agent-action scenarios.
Deliverables: (1) a documented reference architecture, (2) a working demonstration of pre-execution decisioning, and (3) an evaluation harness others can run and extend.
How will this funding be used?
Funding supports the founder's dedicated engineering time over roughly three to four months to build the reference implementation and evaluation harness, plus modest compute/hosting for a public demonstration environment.
At the minimum ($10,000), the work delivers a focused reference implementation and a basic evaluation harness covering a core set of agent-action scenarios. At the full goal ($50,000), the scope expands to a broader scenario library, hardened tamper-evident logging with offline verification, and documentation so external researchers can adopt and extend the harness.
Who is on your team? What's your track record on similar projects?
Boone Carlson, founder and principal of Keystone Digital Holdings LLC, is the sole architect of Synchronicity. Relevant standing in this exact domain:
Named author/reviewer in the OWASP Application AI Security Verification Standard (AISVS) 1.0.
Contributor to OWASP's State of Agentic AI Security and Governance.
Work cited in the NIST COSAiS AI control-overlay effort.
Named inventor across six patent families in pre-execution AI governance.
Synchronicity already exists as a working reference implementation with a live demonstration environment, built and self-funded to date.
What are the most likely causes and outcomes if this project fails?
The most likely failure mode is scope: as a solo effort, the breadth of demonstrated scenarios is bounded. The mitigation is to scope the reference implementation and harness to a focused, representative set of agent actions rather than attempting universal coverage.
Even in the downside case, the public artifacts — reference architecture and evaluation harness — advance the field's ability to reason about external agent governance, and the core properties (determinism, fail-closed behavior, verifiable audit) remain demonstrable at smaller scope.
How much money have you raised in the last 12 months, and from where?
To date, Synchronicity has been entirely self-funded by Keystone Digital Holdings LLC. No external investment or grant funding has been raised.