Outcome Intelligence Protocol · v1.0
The registry for trusted Outcome Intelligence.
OIP is an open protocol for publishing, verifying, and trusting reusable expertise that helps agents and people reach real-world outcomes — with the evidence to back it up.
Public Review Draft · feedback welcome on the review page.
What is Outcome Intelligence?
Agents can call any model and any tool. What they cannot easily answer is which packaged expertise to trust to deliver an outcome — and how they know. Today that judgment is made from proxies that are not evidence: star ratings, download counts, model benchmarks.
OIP is the layer that makes the question answerable. An Outcome Intelligence object is reusable, provenance-backed, versioned expertise aimed at one real-world outcome — and it carries an evidence-led trust signal, not a star rating.
What OIP standardizes
Outcome Manifests
A machine-readable description of what an outcome object does, how to install it, and what it costs.
Trust Manifests
Provenance, freshness, and an evidence-led trust signal — kept separate from marketing copy.
A read API
Discover and evaluate outcomes via /.well-known/oip.json, /v1/outcomes, and /openapi.json.
Compatibility
Complements A2A, MCP, SKILL.md, ACP/UCP, and Apple App Intents. It does not replace them.
Trust is evidence-led, not star-led
Every trust signal is shown with its band, a confidence level, and how much evidence stands behind it — never a naked score. A one-journey result is honestly marked as early evidence:
See the worked example in the registry preview.
Where it stands. OIP v1.0 is a Public Review Draft. SkillSenz operates a reference registry with a single example object. DeepSenz builds the reference implementation. This is not a final standard, and nothing here is a marketplace yet.