Microsoft releases its multi-model Copilot harness through Foundry

The harness is the system around the model

Microsoft is making the agent harness used across its Copilot products available through Microsoft Foundry. A harness is the orchestration around a model: instructions, context, tool access, execution environment, agent loop and governance. Microsoft says the same foundation powers Copilot Code, Copilot Studio, the GitHub Copilot app, the command line and Visual Studio Code.

The platform supports more than 20 frontier models and bring-your-own-key access. It can route work to different models without requiring the application to replace its orchestration layer. Open standards include Model Context Protocol, agent skills and plugins, allowing developers to combine model choice with tools and reusable domain instructions.

GitHub engineering becomes a shared Microsoft layer

The harness originated in GitHub engineering and has been standardised across Microsoft. Its runtime was ported to Rust, which Microsoft says reduced session overhead and time to first token. Sharing improvements across several products can improve consistency, although public claims about performance should be validated on the customer’s own workload and model mix.

Developers can download an SDK or run the harness as a managed Foundry service. In the managed arrangement, Microsoft operates the agent loop and sandbox while the customer supplies instructions, skills, tools and model choice. That division can reduce maintenance, but organisations remain responsible for whether those components grant appropriate authority and produce acceptable outcomes.

Managed routing still needs customer evaluation

Multi-model routing introduces evaluation questions that a single-model application can avoid. A task may move between providers with different retention, regional, safety and pricing characteristics. Teams need to know when routing changes, preserve the model and configuration used for each decision and define which workloads may never leave an approved deployment boundary.

The release reflects competition shifting from individual models to the systems that make them useful. A strong harness can improve tool reliability and reduce repeated engineering. It can also become a concentrated dependency. Buyers should test portability, failure behaviour, observability and the ability to pin a model when reproducibility or regulation matters more than automatic optimisation.

What teams should validate before production

For implementation, teams should record the exact model, software, dataset, connector and policy versions behind every result. Representative evaluation must include normal work, boundary cases, deliberate misuse and component failure, with outcomes separated by user role, data sensitivity and consequence. A strong average should never hide a serious failure in a critical path. Changes should pass the same regression set before promotion, and operators need a documented fallback that preserves evidence rather than silently broadening access or lowering quality.

Procurement and planning should account for the human work required to review exceptions, maintain controls and investigate incidents. Published performance or efficiency claims are useful starting points, but production confidence comes from evidence collected in the organisation’s own environment. A named owner, measurable acceptance threshold, cost model and tested shutdown path turn a promising release into a service that can be governed over time. Where decisions affect safety, rights, money or regulated records, accountable professional review remains essential.

Security and operational evidence

Security review should map every identity, credential, network destination, stored artefact and administrative access path involved in the workflow. Logs need enough context to reconstruct who initiated an action, which evidence was consulted, what the system proposed and what ultimately executed. Sensitive prompts and traces should not be retained merely because they are useful for debugging. Retention, redaction and access must follow the underlying data classification, with a practical process for correcting or removing material when policy or law requires it.

Operational monitoring should measure quality and service behaviour together. Latency, throughput and availability matter, but so do unsupported claims, missed constraints, unsafe tool calls and the rate at which people overturn recommendations. Teams should rehearse dependency outages and partial failures before unattended operation. If a model, connector or external service becomes unavailable, the workflow should fail predictably, preserve its state and tell the user what remains unfinished. Clear degradation is safer than an agent improvising around a missing control.

Transparency should scale with adoption

Organisations should explain the scope and limitations of the deployed capability to the people affected by it. Users need to know when automation is involved, how to challenge an outcome and where a human can intervene. That transparency improves feedback and prevents a pilot result from being interpreted as blanket approval for broader autonomy. Expansion should follow demonstrated evidence, not novelty or pressure to match a competitor announcement. The durable advantage comes from dependable integration, careful measurement and learning that survives the next model or product update.

Finally, teams should schedule a review after real usage rather than treating launch as the end of governance. The review should compare expected benefits with observed quality, cost, incidents and user behaviour, then retire workflows that do not justify their complexity. A service that works well can expand through the same evidence-based process. A service that fails should be narrowed or stopped without losing the records needed to understand why. This feedback loop is what turns experimentation into accountable operational capability.