The protocols · scientific & technical

FDE × DeepSCR — a protocol for verifiable AI engineering.

Two protocols compose into one discipline: DeepSCR is a governance protocol — separation of powers over any work — and FDE is the domain method it governs. Together they turn a vague AI ask into a falsifiable, evidence-grounded, certified deliverable. This page shows the principle, the flow, and a worked engagement.

Follow the flow ↓See certified output
1 · The principle

Governance over a domain — not a template.

DeepSCR is domain-agnostic: it could certify any work. FDE gives it engineering depth. The composition is hierarchical — DeepSCR governs FDE specialists — which is what makes the output auditable.

Governance protocol

DeepSCR

Separation of powers: no single agent both proposes and certifies. Each step is adversarial to the previous one.

  1. S — Hypothesis: state a falsifiable claim.
  2. C — Contradiction: attack it; ≥3 failure modes before shipping.
  3. V — Verification: every claim ties to ground truth (file·line·test).
  4. R — Certification: an independent 0–100 Assurance Score, with veto.
Domain method

FDE — Forward Deployed Engineering

The field method (Palantir lineage) for taking a business problem to owned, production software.

  1. 0 — Reconnaissance: scan the real codebase/business.
  2. 1 — Scoping: the 6-Q decomposition.
  3. 2 — Prototyping: candidate architectures, held-out gate.
  4. 3 — Production: runbook, evals, handoff.
  5. 4 — Feedback: reusable field intelligence.
2 · The flow on a problem

A worked engagement: reduce B2B SaaS churn.

Watch the data flow through the four powers — each stage's output is the next stage's adversary. Nothing advances until it survives.

data-flow · problem → certified deliverable → public registry
Problem6-Q intake HypothesisLead ContradictionResearcher VerificationBuilder + 4 FDE CertificationCertifier Registryhash chain
S

Hypothesis — the Lead decomposes the problem

DeepSCR · 6-Q decomposition
In "Reduce churn with AI" — vague.
Out 6-Q: Q1 process · Q2 classification, <1s, ≥0.80 AUC · Q3 24-mo events, GDPR · Q4 €12k/churn · Q5 manual 55% recall · Q6 −18% churn, 200% ROI. + falsifiable claim + hypothesis tree
C

Contradiction — Researchers falsify in parallel

DeepSCR · held-out gate
In The hypothesis tree.
Out Each hypothesis attacked on a held-out set; only survivors promoted; ≥3 failure modes kept with lessons. promote ⇔ beats the gate
V

Verification — Builder consults 4 FDE specialists

DeepSCR governs FDE · ground truth
In The promoted architecture.
Out Scoping · Architecture · Agent · Production specialists assemble the deliverable; every claim cites a real file·line·test the Certifier resolves on disk.
R

Certification — independent score + veto

DeepSCR · separation of powers
In The deliverable + its evidence trail.
Out FDE Assurance Score 0–100; ship only if ≥85; written to the public hash-chained registry. Otherwise: veto → re-routed to the weakest owner.
3 · The measure

The FDE Assurance Score — falsifiable by construction.

Four weighted components. Evidence is the largest — and it only counts if the cited artifacts resolve on disk (ground truth, not structure).

Claim present & falsifiable25
Contradiction (≥3 failure modes)25
Evidence trail (file·line·test, verified)30
Anti-patterns clean12 / 20
Assurance Score
0/100
✓ CERTIFIED ≥ 85
this engagement: 92 ≥ 85 → ship · a score < 85 → veto, no ship
4 · The team

8 agents — 4 govern, 4 do the domain.

The Builder consults the four FDE specialists in parallel; the Certifier verifies their claims. "DeepSCR governs FDE specialists."

AgentLayerRole in the protocol
LeadDeepSCR6-Q decomposition + falsifiable claim + hypothesis tree
Researcher ×NDeepSCRFalsify hypotheses in parallel against the held-out gate
BuilderDeepSCRAssemble the deliverable from the specialists; score the rubric
CertifierDeepSCRIndependent Assurance Score; veto if evidence is weak
ScopingFDE specialistSharpen the 6-Q; lock metric & cost-of-error
ArchitectureFDE specialistStack / SaaS architecture from Q2–Q3, kept reversible
AgentFDE specialistEval gate, reason codes, guardrails, red-team cases
ProductionFDE specialistRunbook, observability, cost ceiling, compliance
Why it holds up

Four properties that make the output scientific.

FALSIFIABILITY

Claims can be wrong

Every engagement states a claim a held-out test could refute — not a vibe.

SEPARATION

No self-certification

The proposer never certifies. The Certifier can veto the Lead's optimism.

GROUND TRUTH

Evidence resolves on disk

Synthetic structure earns nothing; cited file·line·test must exist.

REPRODUCIBLE

Public & auditable

Every certification lands in a hash-chained registry anyone can re-verify.

From principle to practice

Run the protocol on your repo.

Install the Skill, point it at real code, and read the certified result.

Install the Skill →Browse the Registry