Source QC
Trace originals, preserve context, and mark what a record can—and cannot—establish.
Global quality-control community · human signal archive
A public working archive and coordination layer for quality controllers across cultures, disciplines, and time zones. Read publicly. Join deliberately. Keep the final judgment human.
Open global call
A QC is anyone willing to slow down, separate evidence from inference, preserve uncertainty, and let the record change their mind. No credential, country, discipline, or belief system owns that responsibility.
Trace originals, preserve context, and mark what a record can—and cannot—establish.
Stress-test protocols, failure conditions, controls, and claims of replication.
Carry meaning across languages without laundering certainty or erasing local context.
Notice what distant institutions miss and route grounded knowledge into the record.
Connected record set
These files become useful when read together. Their presence in a government archive establishes institutional attention—not the truth of every claim inside.
Contractor proposal
A $30M blueprint for turning alleged anomalous cognition into a research and intelligence program.
Decode →Army analytical memo
A model of altered attention that combines practical exercises with highly speculative physics.
Decode →Program decision memo
A budget-and-policy snapshot recommending that the existing appropriation be spent as planned.
Decode →Research architecture
7 June 1991 · SAIC → DIA
SAIC proposed a five-year, 175 person-year program costing about $30 million: select and train “receivers,” standardize blind sessions, score responses, test applications, and route any surviving output into ordinary human workflows.
Established by the record
Asserted by its authors
Later operational finding
Gateway decoded
9 June 1983 · U.S. Army Operational Group
Wayne M. McDonnell’s memo analyzes the Monroe Institute’s Gateway Experience. It combines relaxation, sound, visualization, and self-observation with a speculative account of consciousness, holograms, time, and the “Absolute.”
Operational layer
Theory layer
Operational translation
We preserve the exercise, strip away implied certainty, and add a measurable failure condition.
Write down distractions, risks, and expectations before the session. Park them; do not erase them.
Use 60–90 seconds of comfortable paced exhaling or humming. Audio is optional; no medical claim is implied.
Translate “mind awake / body asleep” into a short body scan: physically settled, mentally attentive.
Open attention, then record raw impressions before interpretation. Keep nouns and explanations in a separate field.
State one concrete question in advance. Record the predicted next action and what would count as failure.
Log confidence, mood, expectations, and misses. The observer is part of the measurement system.
The official CIA scan skips report page 25. A Monroe Institute copy reported in 2021 supplies a candidate page, but that recovery has secondary—not CIA—provenance.
Program layer
30 August 1993 · DDR&E → USD(A)
The memo summarizes congressional funding, contractor concentration, DIA staffing, and three continuation choices. It recommends no change: finish the work supported by the existing appropriation.
Congressional-add history
Internet archive finding
Its two-page object is a fax cover plus decision memo. The “full original” program record is not one longer version of this memo; it is the CIA’s STARGATE collection.
The exact two-page record supplied here: fax cover plus decision memo.
The program corpus. Use this—not one memo—as the archive boundary.
A primary methods document for reconstructing what operators were asked to do.
A program-era source for claims, protocols, and internal framing.
The later operational review that found continued intelligence use unwarranted.
Fleet protocol
Recursive learning · 12 minutes per loop
One person protects the target, one records impressions, one attacks the inference, and one preserves the audit trail. Repeat the loop to improve the protocol—not to turn repetition into proof.
Define the question and failure condition.
Label source, observation, and inference.
Keep target information from the observer.
Reveal only after every note is locked.
Change one protocol variable at a time.
Release misses, uncertainty, and provenance.
Local field console
Nothing entered here is uploaded or shared. Copy the locked packet into your community’s own record.
Frame
01:00Source Keeper assigns an opaque target ID and defines the failure condition.
Cross-file synthesis
A training theory and proposed altered-attention model.
A contractor blueprint for selection, testing, and operations.
A funding and continuation decision inside government.
A negative verdict on continued operational intelligence use.
Declassification makes provenance inspectable; it does not validate the claims.
Primary archive
Start with the exact file, then move to adjacent methods, budgets, reviews, and the full program collection. Preserve object type at every step.
CIA-RDP96-00789R003100080001-9Contractor proposalScope + budget →Gateway analysisCIA-RDP96-00788R001700210016-5Army memoModel + exercises →STARGATE summaryCIA-RDP96-00789R002500240008-1Decision memoFunding + options →1995 evaluationCIA-RDP96-00791R000200180005-5Independent reviewOperational verdict →Mirror integrity · 59 pages · 3,636,195 bytes · SHA-256 53b9083a506c…2604 · Full digest and evidence boundaries are published in the JSON manifest.