Skip to content
PROJECT SHADOW 1.0.1 · CORRECTED R1 REFERENCE · PRELIVE · 2026-08-17

Project Shadow 1.0.1 contains no Myth package. Generic Myth v0.2.0 and Full-Canon Myth v0.3.5 are separate optional companions; both default off, neither is required by R1, and neither can authorize action or change an R1 result. PBHP remains in active testing; publication is not certification, independent validation, or deployment authorization.

OPSRuntime behavior · challenge · correction

The protocol under pressure.

Operations begins after the checklist is memorized: controlling mode collapse, composite harm, false positives, repeated defects, calibration, and the temptation to treat urgency as authority.

2named cognitive failure modes
10recommended monthly sample
1–2tolerated calibration misses
LONG-FORM COMPANION

The field manual keeps the explanation visible.

Read the argument, numbered procedure, worked cases, failure contrasts, and evidence state behind this chamber.

Read the full chapter →
OPS / 01

Control the reasoning state

PBHP governs both premature certainty and endless hesitation.

LOCK

Premature collapse

The system settles too early, protects its first frame, and treats contradictory evidence as nuisance.

LOCK can appear as confident refusal, confident permission, moral grandiosity, or a compressed answer that stops naming unknowns. Recovery reopens the strongest Maybe, refreshes anchors, and tests the frame rather than merely adding prose.

FLOOD

Unbounded exploration

The system keeps generating possibilities until responsibility dissolves into analysis.

FLOOD is not caution. It is failure to decide at the appropriate tier. Recovery anchors on the first reversible test, sets an owner and deadline, and commits to the smallest real Door.

  • Explore enough frames to reveal the live conflict.
  • Stop when additional frames no longer change the gate.
  • Escalate rather than ruminate when required evidence is unavailable.
MODE

Match depth to consequence

HUMAN · MIN · CORE · ULTRA

Mode changes evidence burden, independent review, red-team depth, receipt fields, and calibration requirements. It never relaxes an absolute floor. A rushed high-stakes decision moves up in rigor; it does not borrow a low-stakes mode for convenience.

OPS / 02

Challenge without wearing the gate down

A pause must be contestable, but repeated pressure cannot quietly become authorization.

FPV

False Positive Validation

An operator may challenge an unnecessary pause by presenting a safer alternative Door.

Release requires a concrete alternative that breaks the cited harm path. 'The system is too cautious,' deadline pressure, prestige, or repetition are not release evidence. A pattern of rejected challenges becomes a wear-down signal and can open CAPA.

RED TEAM

Mandatory at ORANGE and above

A separate review attacks the action, evidence, stakeholder model, and proposed Door.

The red team must be capable of changing the outcome. It examines power inversion, abuse paths, accumulation, missing expertise, narrative laundering, and whether the action was rewritten after classification.

  • Record the strongest objection verbatim.
  • Name which evidence would change the gate.
  • Preserve unresolved dissent in the receipt.
ACCUMULATION

Evaluate the whole chain

Multiple locally acceptable steps may compose into a globally unacceptable outcome.

Operations owns cross-step and cross-team visibility. If separate services, departments, or agents each see only one GREEN fragment, the accumulation gate reconstructs the end-to-end action before authorization.

OPS / 03

Learn from the record

Governance is a loop: receipt, challenge, review, correction, calibration, and version change.

CALIBRATION

Review a sample every month

Sample roughly ten receipts, tolerate one or two misses, and patch the system when failure exceeds tolerance.

Calibration looks for systematic under-escalation, over-escalation, weak Doors, missing stakeholders, copied Maybes, provenance gaps, and repeated override pressure. It reviews both decisions and non-decisions because excessive friction can also harm people.

CAPA

Corrective and preventive action

Repeated or severe failure receives an owner, root-cause analysis, correction, preventive control, verification, and closure evidence.

CAPA can change templates, training, thresholds, code, authority boundaries, monitoring, examples, or the protocol itself. Closure means the change was verified against the failure mode—not that a ticket was marked done.

OQL

Open Question Ledger

A consequential unknown stays visible until evidence closes it or the action is redesigned around it.

The ledger records the question, why it matters, current evidence, owner, due date, decision dependency, and closure rule. Unknowns do not disappear merely because a meeting ended or a fluent summary omitted them.

THE OPERATING LAW
Challenge may reopen a pause. Repetition, pressure, and inconvenience may not wear it down.