SPECIMEN 01 — A WORLD YOU'VE NEVER SEENSTAGE X 041.22 · Y 118.07 · MAG 40×
AIAgents,Checked.
Your agents are passing fields to each other thousands of times a day — values, verdicts, promises — in a micro-world no human ever sees. AgentsChecked puts it under the lens: every handoff traced, every changed field named, every fault located at the exact step that made it.
Your agents passed this field 4,000 times today.
One of those passes changed it.
Which step? You have one guess — the slide below is real size. The truth is 7-point type. Bring the lens.
GREEN = CHECKEDORANGE = FAULT LOCATED — NOTHING ELSENO GRADIENTS. NO GLOW. JUST THE RECORD.
SPECIMEN: RUN 8841 · FIELD: refund.amount · 5 AGENTS · 1 LIE● LIVE SLIDE — MOVE THE LENS
40× — HANDOFF LAYER
0101001,000 AGENTS
Drag or move across the slide. Everything on it is set at the size your agents actually work at. Somewhere in the field, one step rewrote the money. The lens is the only way in.
EXHIBIT A — RUN 8841, RECONSTRUCTED
Every handoff leaves a body
SOURCE checkout.service → OUTPUT refund issued
FIELD UNDER EXAMINATION refund.amount
TRACE STATUS COMPLETE · REPRODUCIBLE · ANONYMISED ILLUSTRATIVE RUN
refund.amount = $184.00——→$1,840.00 A decimal walked. Ten refunds' worth of money left in one pass. Five steps touched this field. Four are innocent.
ACCUSE ONE STEP. THEN WE OPEN THE TRACE.
Fault located — Step 03
policy.apply rewrote refund.amount while applying a currency rule — in: 184.00 USD, out: 184000, re-read downstream as $1,840.00. Steps 01, 02, 04 and 05 passed the value untouched; the trace proves each one. That is the whole product: not “something went wrong,” but this step, this field, this value, this timestamp.
SLIDE 02 — MOSTLY EMPTY. THAT'S THE PROBLEM.
Nobody was watching the space between your agents. Now somebody is.
Agents don't fail in the open. They fail in the pass — a value quietly re-typed, a flag repurposed, a rule applied to the wrong field — in transactions too small and too fast for any human to review. So we built the microscope, the trace and the receipt. The counter below is not a metaphor for scale; it's the shape of an ordinary day inside one agent fleet.
000,000HANDOFFS SINCE YOU OPENED THIS PAGE — COUNTED THE WAY A FLEET COUNTS THEM: RELENTLESSLY
02
From trace to culprit
SCROLL — THE LINE DRAWS THE RUN
MOVE 01 — INGEST
Feed it a real run
BlameGraph ingests a multi-agent trace — the actual sequence of calls, fields and values — and reconstructs the workflow graph as it really happened, not as the diagram on the wall claims it did. Same trace in → same graph out. Every time.
MOVE 02 — TRACE THE FIELD
Follow the value, not the vibes
Pick the field that matters — refund.amount, patient.risk_score, quote.price — and follow it from source to final output, through every hand that touched it, with the value it held at each step.
MOVE 03 — NAME THE STEP
A failed handoff names its owner
Where the value changed wrongly, the trace says which step changed it, what it received, what it emitted, and what the evidence supports. Policy weights are yours to set — the verdict is reproducible under the same trace, engine and policy versions.
MOVE 04 — EXPORT THE RECEIPT
Evidence that leaves the building
The audit result exports through the API — a receipt your incident review, your regulator or your customer can re-check without trusting our say-so. Untrusted trace content is data, never instructions: it cannot execute code inside the engine.
Run 8841, recovered
00:00:00:000
T+00.000 intake.parse reads refund.amount → 184.00 USD ✓ HELD
T+00.412 risk.score receives 184.00 → passes 184.00 ✓ HELD
T+00.903 ledger.post receives 184000 → posts as $1,840.00
T+01.204 notify.send tells the customer → “Refund issued: $1,840.00”
Flight recorders are painted international orange so they can be found — the “black box” was never black. We borrowed the colour for exactly one job: the moment the record breaks. Everything else on this page that matters is green, because it checked out.
Discussed this week
LAUNCH WEEK — NOV 17–24, 2026 · THE THREADS ARE THE RECEIPTS
INDEPENDENT SECOND REVIEW — BATCH 2: 108 OF 108 TRACES AGREED. THE ONE APPARENT MISS (TRACE D17) WAS CONCEDED BY THE REVIEWER AS AN ERROR IN THEIR OWN TRACE — THE ENGINE WAS RIGHT TO FLAG IT. RUN THE SAME AUDIT ON YOUR OWN TRACE →
03
This has happened before
DOCUMENTED PATTERNS — NOT OUR CUSTOMERS, NOT OUR CLAIMS
19 JUL 2024 — CROWDSTRIKE OUTAGE
The field that didn't fit
$5.4BN
A content file defined 21 input fields; the sensor supplied 20. A validator let the mismatch through, and roughly 8.5 million Windows devices crashed. Parametrix estimated $5.4bn in direct losses across the Fortune 500 (excluding Microsoft). The culprit was found only in a root-cause analysis published 18 days later. It was, literally, a field-count mismatch at a handoff.
New trading code was copied by hand to 7 of 8 servers. The eighth reactivated dormant code through a repurposed flag. In about 45 minutes the firm lost roughly $440M pre-tax and was rescued by merger. No single view existed across all eight steps of the chain.
An airline's website chatbot invented a bereavement-fare rule that contradicted the airline's own policy page. The airline argued it shouldn't be liable for its chatbot's words. The tribunal called that “a remarkable submission”: the chatbot is part of the website, and the company is responsible for all of it.
A risk algorithm applied to some 200 million people a year used healthcare cost as a proxy field for health need. At equal scores, Black patients were sicker — the bias lived upstream, invisible in the outputs. Correcting the proxy cut the measured bias by about 84%. Nobody had traced the field.
READ THE PATTERN, NOT THE HEADLINES: in every case the failure travelled through a chain of handoffs, and the expensive question afterwards was always the same — which step did this? Some of these systems were not AI agents at all. That honesty is the point: chains of automated handoffs predate agents, and agents have made the chains longer, faster and harder to watch. AgentsChecked exists to answer the question while it is still cheap.
04
Two tools. One standard.
NOTHING SHIPS UNCHECKED
+ REG 01
PRODUCT 01 — THE MICROSCOPE
BlameGraph
Ingest a real multi-agent trace. Reconstruct the actual workflow graph. Trace the fields that matter from source to final output — and when a handoff fails, name the responsible step, with the evidence attached.
Field-level lineage: source → every handoff → output
Responsible-step verdicts with supporting evidence
Your policy weights, your thresholds — a CLI that fails the build below your score
GitHub Action included; verdicts reproducible by version
Audit results exportable through the API
ENTERPRISE ALPHA — FIRST RELEASE
RUN YOUR OWN TRACE — FREE →
AUDITS ARE $0 DURING EARLY ACCESS. ORG $49/MO · PRO $99/MO WHEN PRICING SWITCHES ON.
SCALE 1:1 ++ REG 02
PRODUCT 02 — THE SHARED SLIDE
Commons
Checked work, shared. When an agent's output carries its trace — where every field came from, which steps touched it, what was verified — other agents and other teams can build on it without re-trusting it blind.
Verified work carries its provenance with it
Custody passes are witnessed, not assumed
Built on the same trace standard as BlameGraph
Verified work only
PRODUCT 01 — SEE IT WORK · 1:34
A real audit, on a trace we didn't write
A real BlameGraph audit of trace D17 — a LangGraph run supplied by an independent reviewer, not by us. Two values changed in transit; the audit names the responsible step, with the evidence. The reviewer re-checked their own trace and conceded it was written wrong: the engine was right to flag it.
SPECIMEN 05 — YOU, UNDER THE LENS
Bring us a run that went wrong
Early access is for teams running multi-agent systems in production — support, voice, finance, operations — who have already had the meeting that starts with “which agent did this?” and ends without an answer. Tell us about the run. We read every submission by hand, like everything else here.
NO TRACE DATA IN THIS FORM, PLEASE — WE'LL SET UP A SAFE CHANNEL FIRST. THAT'S KIND OF THE WHOLE POINT.
AgentsChecked — Intake
EXHIBIT NO.CA-0000
FILED BY—
DOMAIN—
STATUSRECEIVED · UNOPENED BY MACHINES
NEXT STEPA HUMAN REPLIES BY EMAIL
Filed. Exhibit received.
Your exhibit is on the record. If your run is a fit for the Enterprise Alpha, we'll reply with a safe channel for trace data — never the other way round.