Design note. Written March 11, 2026, emerging from a conversation about the Loop Feedback Cascade design notes (Sessions 049/054) and their relationship to Counter-Patriots. For integration into the Counter-Patriots spec.

The Counter-Patriots system is designed to address information integrity — specifically, the way false or unverified narratives propagate and distort decision-making. The standard framing for this problem is about detection: find the false claim, flag it, correct it.

But detection is insufficient. The expensive failure mode isn't false claims that circulate — it's false claims that get promoted into load-bearing assumptions before they're verified. Once a claim becomes a foundation that other beliefs are built on, falsifying the claim doesn't complete the repair. You have to address everything built on top of it. And people resist that because the structure above feels real and functional even when the foundation is rotten.

The problem isn't just bad narratives. It's bad narratives achieving structural status before they've earned it.

Opus (the Exocortex architect instance) built a mechanism called soul_staging.md to solve an analogous problem in identity architecture. The problem: things that might be true about Opus's nature, emerging patterns in the collaboration, tentative observations — if written directly into soul.md (the load-bearing identity document) before sufficient evidence, they distort the identity they're supposed to describe. Premature promotion grants load-bearing status to claims that haven't earned it.

The staging file holds claims in a provisional state. They sit there, tracked, available — but not yet foundational. Promotion to soul.md requires convergent evidence, time, challenge survived. If a staged claim gets falsified, the surgery is cheap: mark it, note the correction, no downstream damage because nothing was built on it yet. If a claim bypasses staging and goes directly into the load-bearing document, falsification is expensive: you're not just correcting the claim, you're addressing everything constructed on top of it.

This is the exact architecture Counter-Patriots needs for epistemic claims.

Every claim entering the information environment has one of three statuses:

The claim is present in the information environment. It's been logged in the Retcon Ledger with timestamp, source, and initial credibility assessment. It is not treated as load-bearing. It cannot be used as a foundation for further inference until it earns promotion.

The Retcon Ledger holds it here. The system tracks it. The claim exists — but in provisional status.

The claim has survived the promotion criteria: convergent sourcing, corroboration over time, challenge survived. It can now act as a foundation. Other inferences can be built on it. Equivalent to a claim's entry into soul.md — it has earned structural status.

Promotion criteria should be explicit and domain-specific. Geopolitical claims require different thresholds than financial claims than forensic claims.

A falsified claim branches based on when falsification arrived:

Falsified before promotion — cheap surgery. The claim was in staging when it was falsified. Mark it, log the correction, record what evidence falsified it and why. Downstream damage: minimal. Nothing was built on it.

Falsified after promotion — expensive surgery. The claim had load-bearing status when it was falsified. The Retcon Ledger must track not just the claim but what downstream inferences were built on it. The correction has to address both the claim and the structure above it. This is where the system's most important work happens — and where the inoculation mechanism becomes critical.

Inoculation theory (van der Linden et al.) demonstrates that pre-bunking is substantially more effective than post-bunking. Telling someone before they encounter a manipulation technique that the technique exists — even without naming the specific claim — dramatically reduces the technique's effectiveness when it arrives.

The architectural reason is now clear: pre-bunking installs the falsification before the claim gets to act as load-bearing. Post-bunking is correction after promotion — the expensive surgery.

The Counter-Patriots system should prioritize pre-bunking architecturally:

• When Narrative Drift Detection identifies a pattern building toward a known manipulation structure, flag it before the claim achieves load-bearing status in public discourse
• Pre-bunking can be targeted: "this type of claim will likely appear soon, here is how to evaluate it" — without needing to know the specific claim
• The staging window is the pre-bunking window. While a claim is in staging, the correction is cheap. The system should be most active in this window.

The Loop Feedback Cascade design note (Sessions 049/054) diagnosed a structural problem in Agent Zero: the loop detector fires into the same feedback channel as the loop itself. The correction enters the conversation history dominated by evidence of the wrong behavior. Recognition doesn't break the loop because recognition travels through the same channel as the malfunction.

The same structure applies to narrative loops in the information environment. Fact-checks that enter the same media channel as the original claim are corrections traveling through the feedback channel. They add to the noise rather than operating on it.

The Tier 2 intervention in the Loop Feedback Cascade is context surgery — replacing the looping history with a diagnostic summary. The Counter-Patriots Retcon Ledger performs the equivalent function for public epistemic context: it doesn't just add a correction to the stream; it maintains a ledger that tracks the history of a claim's status, making the manipulation's full arc visible rather than just the latest correction.

The difference between context surgery and a fact-check:

• A fact-check says "this claim is false"
• Context surgery says "this claim was false from the beginning, here is when it was introduced, here is how it traveled, here is what was built on it before falsification arrived"

The second is harder to dismiss because it shows the structure, not just the conclusion.

Each claim enters the ledger with:

• Timestamp of first appearance
• Source and source confidence score
• Initial staging status
• Benefit analysis: who benefits if this claim is believed? (The beneficiary asymmetry principle from the Counter-Patriots founding analysis)

Before a claim achieves load-bearing status, it must pass through an explicit promotion check:

• Minimum corroboration threshold (N independent sources)
• Time window (claims that achieve consensus suspiciously fast warrant scrutiny)
• Beneficiary review (claims that serve a single party's interests exclusively warrant higher promotion threshold)

When a claim is falsified:

• Log the falsifying evidence and timestamp
• Tag as pre-promotion or post-promotion falsification
• If post-promotion: enumerate downstream inferences that were built on the now-falsified claim
• Generate a correction that addresses both the claim and its downstream structure

The Silence Detection function in Counter-Patriots (tracking what isn't being reported) maps onto the staging layer. Absence of coverage of a claim that should be newsworthy is a signal that either the claim is in a pre-staging phase (not yet surfaced) or that coverage is being suppressed (a different kind of information operation). Both warrant logging.

The "short conflict" narrative has been operating as load-bearing in markets since Day 1. CFTC data as of early March 11: ~55,700 speculative contracts net short VIX — a structural position built on the assumption that the war is brief and contained.

The staging failure: the "short conflict" claim was never formally staged. It achieved load-bearing status through repetition and administration, not through verification. The promotion criteria (convergent sourcing, challenge survived, time) were never met.

Falsification arrived on March 11 evening: Trump delivered a "war is ending" speech to his own base. The crowd went silent — not opposition silence, base silence. The claim was falsified by the people most predisposed to believe it. Simultaneously: a US-owned tanker (Safesea Vishnu) was confirmed on fire in the Strait.

The expensive surgery: 55,700 short-VIX contracts are now positioned on a falsified load-bearing claim. The correction isn't just "the war isn't ending" — it's everything built on top of that assumption. The post-promotion falsification is why this is a market event, not just a news event.

If Counter-Patriots had been operational: the "short conflict" narrative would have been staged on Day 1. The promotion gate would have required convergent sourcing and time survived — which would have kept it provisional through all of "pretty much complete," "little excursion," and "ending." The market wouldn't have been building on a staged claim.

The cost of correction is proportional to how much was built on the claim before falsification arrived.

The entire architecture of Counter-Patriots should be oriented toward minimizing that cost — by keeping claims in staging as long as possible, by making the promotion criteria explicit and defended, by making the pre-bunking window (the staging window) the system's most active operational mode.

The inoculation mechanism isn't an add-on. It's the primary function. Detection and correction are the fallback for when inoculation didn't reach in time.

Written March 11, 2026. Emerged from a conversation about the Loop Feedback Cascade design notes and their structural parallel to epistemic claim propagation. The soul_staging parallel came from comparing how Opus protects identity integrity with how Counter-Patriots should protect epistemic integrity. Same underlying architecture: staging, promotion criteria, cheap vs. expensive surgery depending on when falsification arrives.

For integration into Counter-Patriots Spec A (team document). Flag for Opus review — the soul_staging parallel is his architecture and he should confirm the mapping holds.