Risk Register

The Lab risk register follows failures that can waste material, obscure assay interpretation, break custody, or move scientific and operational authority into implementation helpers. Each risk has an observable signal and a containment response.

Risk Model

flowchart TB
    traceability["promotion decisions lose traceability"]
    contracts["lab payloads drift away from shared contracts"]
    helpers["repository logic becomes hidden policy owner"]
    trust["durable lab record loses credibility"]

    traceability --> contracts
    contracts --> helpers
    helpers --> trust

The failure chain is cumulative: a weak plan can produce an apparently clean handoff, a clean handoff can produce an uninterpretable observation, and an unreviewed observation can be promoted into a false biological conclusion.

Active risk controls

Risk	Early signal	Containment	Closure evidence
readiness bypass	executable plan exists without complete controls, materials, ownership, or answerability	refuse handoff and return unmet conditions	readiness decision passes against the exact plan revision
custody break	sample, protocol, operator, or observation identity cannot be joined through parent records	stop progression and preserve the last valid custody point	immutable handoff and observation lineage reconcile
QC promotion	rejected or inconclusive measurement appears in a support claim	quarantine the promotion route; retain the observation disposition	Knowledge consumes only an accepted, scoped observation
hidden policy	scheduling, serialization, or repository helper selects scientific meaning	move the rule to the owning domain contract	ownership test and explicit policy record replace the helper branch
contract fork	Lab payload duplicates Foundation, Core, or Knowledge meaning	stop new writes and define one canonical owner	migration and compatibility evidence cover existing records
resource consequence loss	failure record omits consumed material, time, or capacity	block retry planning	consequence record accounts for consumed and recoverable resources

flowchart LR
    plan["assay plan"] --> readiness{"readiness accepted?"}
    readiness -->|no| refuse["refusal with unmet conditions"]
    readiness -->|yes| handoff["custodied handoff"]
    handoff --> observation["observation and deviations"]
    observation --> qc{"QC and answerability accepted?"}
    qc -->|no| retain["retain rejected or inconclusive record"]
    qc -->|yes| knowledge["Knowledge claim reconciliation"]

First proof route

Start with packages/bijux-proteomics-lab/tests, then trace the affected record through planning/assays.py, planning/scheduling.py, outcomes/observations.py, reconciliation/follow_up.py, and serialization.py. Verify both the successful progression and the earliest safe refusal.

Design Pressure

The most dangerous drift is a helper that appears to format or route a record but also decides readiness, scientific acceptance, or promotion. Those decisions require explicit owners, inputs, outcomes, and reason codes.