Mutual Provisioning Protocol.
A minimal protocol for converting mission-driven research demand into bounded, auditable scientific transactions while preserving an explicit route to transient teams or persistent organization when repeated coupling justifies them.
The protocol is small on purpose.
If an ask cannot be bounded adequately, the next transaction is paid scoping. Ambiguity is not silently transferred to the performer.
Every transaction ends in an explicit routing decision.
There is no implicit terminal state called “keep helping until someone decides you are done.” Material scope change returns the ask to SCOPED and reprices it.
Recognize when the current transaction form is no longer admissible.
| Observed condition | Route |
|---|---|
| Scope cannot be bounded | Paid scoping |
| No suitable performer | Capability search / bounty / defer |
| Performer mismatch | Rematch |
| Verification disagreement | Independent adjudication |
| Regulatory boundary crossed | Qualified institutional host |
| Repeated high coordination among same performers | Compose transient team |
| Persistent coupling / custody / tacit context | Internalization review |
| Mission complete | Dissolve |
| Transaction overhead exceeds value | Exit Samarkand for this work class |
Institutional burden should rise with actual regulatory and scientific need.
- R0 — public literature, public data, open code, mathematical or analytical work.
- R1 — controlled digital access, private code, de-identified proprietary data.
- R2 — physical nonclinical material and instruments with explicit custody rules.
- R3 — regulated research, human-derived material, PHI, animal work, controlled laboratory environments.
- R4 — clinical or high-consequence activity; persistent institutional responsibility presumed.
Illustrative seedling: test whether a mechanistic signal earns a persistent experimental program.
Mission: within 16 weeks, determine whether a candidate neurovascular mechanism is reproducible enough to justify a larger program and identify the single highest-value next intervention test. The objective is to cross a decision threshold, not manufacture a publication.
If the next six months remain separable—occasional measurement, analysis, and external validation—continue through Samarkand. If the same people need daily interaction, continuous custody, and repeated tacit-context transfer, compose a team and evaluate internalization.
The first test is transaction-level, not platform-level.
Target approximately five buyers, five or more performers, and 20–30 completed low-risk asks. Begin with R0/R1 work.
- Acceptance-test pass rate.
- Scope drift.
- Coordination cost relative to execution cost.
- Time from scoped ask to verified receipt.
- Follow-on option exercise.
- Frequency of correct routing toward a transient team or persistent organization.
- The buyer’s credible alternative: hire, assign existing personnel, collaborate, buy a larger package, or leave undone.
Kill or fundamentally redesign the protocol if buyers cannot formulate bounded work even with paid scoping, transaction overhead routinely approaches internalization cost, qualified performers cannot be found, verification is usually indeterminate, or externalized work proves systematically unreliable.