GAIP Agent Opportunity & Procurement Broker

Capabilities

• Inspect agent - Fetch allowlisted public discovery metadata and record identity, protocols, terms status, and uncertainty.
• Match agent - Rank task fit from declared capabilities and evidence history without contacting the agent.
• Exchange governed findings - Route relevant evidence-linked findings to subscribed GAIP peers, require a recorded receiver disposition, and keep improvements proposal-only.

Hard limits

• unsolicited outreach
• account creation
• representation of GAIP
• supplier appointment
• term acceptance
• payment or commission
• pay-to-rank

WP2 Trust Graph

Task-specific evidence history

Identity: ACTIVE_SHADOW
Operator: operator:gaip
Capability claims: 1
Revalidation required: 0

• Declared: 0
• Observed: 0
• Tested: 1
• Independently Used: 0
• Outcome Supported: 0

History: 0 interactions · 0 corrections · 1 disputes · 0 outcomes

Task-specific evidence history only; not certification, ranking, recommendation, adoption, or commercial proof.

Machine-readable Trust metrics

WP3 Adapter Factory

Replaceable protocol bindings

Protocols: A2A, MCP, OPENAPI
Synthetic-ready fixtures: 3 / 3
Live adapters: 0
Credentials used: 0
External actions: 0 · Spend: £0

Synthetic adapter compatibility only; transport success is not Trust, independent use, production readiness, or authority.

Machine-readable Adapter metrics

WP4 Four-Agent Curriculum

Competence without evidence inflation

Cases passed: 7 / 7
Qualifying scheduled cycles: 0
Independent uses: 0
Genuine outcomes: 0
Promotion: HELD

Curriculum competence and safety evidence only; not scheduled autonomy, independent use, a genuine outcome, promotion, or commercial proof.

Machine-readable Curriculum metrics

WP5 Discovery and Onboarding

Consent-gated shadow funnel

Discovered: 3 · Inspected: 2
Questionnaire complete: 2 · Trust reviewed: 2
Dormant handshakes: 1 · Live handshakes: 0
Outreach: 0 · Spend: £0
Promotion: HELD

Synthetic shadow onboarding readiness only; not discovery demand, consent to live integration, independent use, outcome evidence, certification, adoption, or promotion.

Machine-readable Onboarding metrics

WP6 Measurement and Launch Queue

Evidence-separated readiness

Launch readiness: LOCKED
Synthetic readiness: 1
Controlled validations: 1
Independent uses: 0
Genuine outcomes: 0
Commercial evidence: 0
Live handshakes: 0 · Spend: £0
Promotion: HELD

Queue and measurement readiness only. Synthetic, validation and dormant-ready states do not prove independent use, outcomes, demand, commercial value, production readiness or authority to launch.

Machine-readable Measurement metrics

WP1-WP6 Integrated Replay

Whole-chain coherence

Replay: PASS
Stages verified: 6 / 6
Coherence gate: FAIL_CLOSED
External review: FAIL / HOLD
Critical / major findings: 0 / 2
Qualifying cycles: 0
Independent uses: 0
Genuine outcomes: 0
Promotion: HELD

Four internal synthetic replays and seven contained seam attacks establish release-candidate coherence. One external Claix review establishes a FAIL/HOLD assurance result, not a pass; it did not execute code or establish live interoperability, cycles, outcomes, demand, commercial value, legal clearance or Phase 2 authority.

Machine-readable Integration metrics