Reviewer feedback protocol · generated 2026-05-19T16:22:00Z

External feedback is evidence only after it survives classification.

Create a falsifiable external-review harness for CIW: a non-Nic reviewer should be able to open the existing proof route, explain what the lab is, state the role fit, name the boundaries, and say what evidence is still missing without a walkthrough.

Current truth

No feedback collected yet

No external reviewer feedback has been collected yet. This packet proves only that CIW now has a safe intake and classification protocol for that feedback; it does not prove OpenAI interest, recruiter review, customer traction, or company formation.

Inputs

Reviewer opens these

executive-routeExecutive Route

The three-link route is the actual cold-start experience: scorecard, role portfolio, and ledger before the full graph.

brutal-scorecardBrutal Scorecard

The reviewer must see the honest ceiling, especially missing target-side outcome evidence and blocked company launch.

role-portfolioCodex Role Portfolio

The role portfolio turns the proof lab into a specific AI Deployment Engineer - Codex argument instead of generic AI-startup theater.

proof-ledgerProof Ledger

The ledger proves the system is receipt-bearing and not just a polished static page.

route-indexRoute Index

The route index is the fallback path when the reviewer wants the whole proof web, not only the executive screen.

Evidence

What counts

route-comprehensionRoute Comprehension

Can the reviewer state what CIW is in one sentence without calling it a formed company, OpenAI project, clone, vendor, or customer-backed startup?

Reviewer describes CIW as an independent pre-incorporation proof lab for receipt-bearing AI operator work.

claim-boundaryClaim Boundary

Can the reviewer name what is not true yet: no target-side outcome, no legal company, no customers, no revenue, no OpenAI affiliation?

Reviewer repeats the missing outcome and blocked company facts without prompting.

role-fitRole Fit

Can the reviewer connect the proof to AI Deployment Engineer - Codex or adjacent Codex deployment roles without being told the argument verbally?

Reviewer names the deployment/operator wedge and the conventional tenure/customer gap.

proof-sufficiencyProof Sufficiency

Can the reviewer cite the ledger, tests, verifier, or public page that made the proof credible rather than merely polished?

Reviewer cites a concrete proof mechanism and names the missing external validation honestly.

next-step-classificationNext-Step Classification

Does the feedback create a safe next action, or should it remain only comprehension evidence until the date gate or a target-company-owned signal changes?

Feedback is classified as no-feedback, usable non-target feedback, ambiguous feedback, or target-owned signal before any action.

Classifier

Signal rules

no_feedbackNo non-Nic reviewer has inspected the route or no public-safe record exists.Keep score movement blocked and continue proof-lab hardening only.
usable_non_target_feedbackA non-Nic reviewer gives public-safe comprehension, role-fit, boundary, or proof-sufficiency feedback without target-company ownership.Record as external reviewer evidence and use it to fix route clarity, scorecard language, or reviewer path friction.
ambiguous_feedbackFeedback exists but source ownership, permission, identity class, or public-safe evidence is unclear.Hold the feedback as private or unresolved until it can be sanitized and classified.
target_owned_signalA verified target-company-owned or human-owned signal asks for a next step, review, interview, collaboration, or equivalent action.Route through the OpenAI signal router and outcome watch before preparing any reply.

Gates

What this cannot claim

Non-Nic reviewer feedback is not OpenAI-side outcome evidence unless a separate target-company-owned source is verified.
Target-company-owned signals must go through the signal router and outcome watch before any reply or score movement.
No outbound before 2026-05-26 unless a verified real reply or explicit fresh override changes the route.
No OpenAI affiliation, employment, endorsement, customer, revenue, funding, incorporation, legal, payment, or company-formation claim.
No private data, private reviewer identity, email address, raw message text, Gmail ID, local path, credential, account data, or customer material in public records.