Published , by The Key editors.
NewCore emerges from stealth with $66M
In short: NewCore came out of stealth on 15 June 2026 with $66M. It describes itself as an agent governance layer that starts alongside an existing identity platform. The launch included an Agentic Skill package for Claude Code, Codex and Cursor, aimed at coding agents.
NewCore announced its launch on 15 June 2026 with $66M in funding. Its public material describes three platform functions: discover, secure and govern. The distinction matters when assessing which part of an agent access problem a product addresses.
Discovery uses integrations that only read data, without requiring teams to register agents first. Our recommendation is to begin an evaluation with the agents a team already runs. Ask which agents discovery identifies and which remain outside its view. Treat that question as an evaluation step, not a claim about coverage.
The secure layer replaces standing credentials with access limited by task and duration. Every agent tool call passes through a gateway. The gateway identifies both the agent and the person behind it. For buyers, our editorial priority is to test those two identities together rather than reviewing an agent name alone.
The govern layer uses one policy engine for people and agents. An action can be allowed, require approval or be denied. Risky calls go to a named approver, who confirms with device biometrics. We recommend testing all three outcomes against a concrete task before deciding whether the policy model fits.
NewCore calls its initial deployment path “Start alongside your identity platform”. Under that model, the existing identity platform continues to handle employee sign in. This is an important boundary for readers comparing an identity platform with an agent governance layer. Our recommendation is to document that boundary explicitly during a review.
The launch also introduced an Agentic Skill package for Claude Code, Codex and Cursor. The package targets coding agents. The existing note highlights their relevance because coding agents act for developers with real production access. We recommend using a representative development task when examining how an agent, its human owner and an approval decision connect.
For readers using the key, NewCore appears in results covering approval for individual actions, MCP policy for each call and evidence for auditors. Those are separate requirements, even when one evaluation considers all three. Our recommendation is to select the required result before comparing product descriptions. A gateway requirement should not become an approval requirement merely because both appear in the same platform.
The funding announcement establishes the launch context, while the platform material describes the proposed controls. We recommend keeping those evidence types separate. Review discovery, credentials, gateway identity and approval as distinct questions. That approach keeps the assessment tied to documented functions rather than the size of the funding round.
Sources: NewCore emerges from stealth with $66M; NewCore: Identity foundation; NewCore: Govern, read 2026-11-03