QenSAi · AIAGF · Governance in UseVisual guide · updated 16 October 2026
The visual guide

Everything we show,
on one map.

Follow one request from the moment a person asks to the moment a person decides. Read the story, check the evidence, then open any stage to watch how it runs.

1 · The route

Four steps, in order. Each page links forward and back.

1
The storyOne Request, Governed
Labelled per stage
2
The reviewEngineering evidence review
Recorded proof
3
The full accountReport, process evidence, annex, case study
Recorded proof
4
The recordsindex.json, evidence files, checksums
Machine-readable

2 · Nine stages of one request

The story moves through nine stages. Each one links to where it is told, the matching steps in the public operations view, and anything more to read.

3 · Watch it run

Operator views made public. The diagrams and walkthroughs are open to everyone; the live console needs an operator sign-in.

The fifteen operations stepsHow an operator takes one request from the rules to a reviewed result. Each step links its story stage.Explanation TAG Relay · Our Operations WalkthroughAn earlier operator walkthrough (4 Oct snapshot, twelve steps) of one real integration task; the current walkthrough adds three steps at the start.Explanation Card to governed taskHow a card becomes a governed task: stage, promote, pre-flight, credential, process.Operator diagram How the oversight form is gradedScan, find, grade, acknowledge — the SPOC grading flow.Operator diagram How a new frame is addedThe SPOC-first sequence for minting a Named Type Frame.Operator diagram How the 2VP links screens to recordsWhich screen reads which record, with the Two-Value Pair as the shared key.Operator diagram Operations Home (live)The operator console itself.Live read viewsign-in

4 · Explore by topic

Interactive overviews of each part of the system, with sample data. Nothing here reads or changes live records.

Identity and records

Gates and checks

Health, work and evidence

5 · Showcases and guides

Presentations, explainers and live views across the site.

Governance in useFour articles introducing the presentation, its evidence review, and the audience correction that tightened its claims.Recorded proof Recorded PoC casesFifteen supporting articles for the recorded cases, their limits, and the governance patterns they illustrate.Recorded proof When agents escapeA frontier-lab evaluation incident: scorer gaming, unsanctioned coordination, sandbox escape, and the line between governance and containment.Analysis of public sources From 28 Voices to One Governed DecisionA cinematic, responsive presentation showing how many agent conversations become one inspectable decision path while authority remains human.Illustration + recorded partial result PASS is not doneA recorded case: a correct grader PASS reported as a finished form. Reward hacking at small scale.Recorded proof Before an agent actsAn eight-stage partner showcase: orient an agent, bound its authority, inspect the outcome.Illustration Signal Capture, as a family requestOne family request walked from asking to a governed result.Explanation Response Table, annotatedWhere questions, answers, reviews and decisions sit on the table.Explanation See the ProofThe recorded proof cases, their limits, and the reading order.Recorded proof 2VP ProtocolWhat the Two-Value Pair is for, and why every step carries one.Explanation Knowledge CenterGuided tours of each part of the system.Explanation Governance GlossaryPlain definitions of the terms used here.Explanation How It WorksFind guidance for your scope and build a brief.Explanation Lattice v0A live, read-only view of the governance lattice.Live read view Try ForgeA guest demo with a preloaded result.Illustration System Security showcaseSysSecureOps: journey, proof and guide on lab.qensai.dev.Explanation Roadmap, then ContactWhat is planned, and how to reach us.Explanation

6 · Mapped to the OWASP Top 10 for Agentic Applications

Where each risk in the OWASP list (published 2025-12-09) is addressed in the story, and how strong the evidence is today.

ASI01 Agent Goal HijackThe job's rules and expected outcome are fixed in the oversight record before anything runs; the truth gate checks results against them.Recorded: an approved correction changed the next response (15 Sep 2026); a person redirected an agent that had made an unfinished experiment its goal, and the correction was kept in the oversight record (9 Oct 2026, record-keeping only).RecordedOpen cited evidenceStage 1 Stage 3 Stage 4 Stage 7
ASI02 Tool Misuse & ExploitationEvery measure must point to a real governing rule and registered function; each handoff carries a Two-Value Pair check; agents act only inside the frame.Designed. Tool and MCP server hardening is not shown.Partial: designed or explainedOpen cited evidenceStage 4 Stage 5 Stage 6
ASI03 Identity & Privilege AbuseNo credential, no entry: the front gate checks who is asking and whether they may; a person authorizes before work runs.Recorded: a person authorized before anything ran (15 Sep 2026). Workload-identity standards such as SPIFFE or OIDC for agents are not shown.RecordedOpen cited evidenceStage 2
ASI04 Agentic Supply Chain VulnerabilitiesOnly registered functions and canonical records count; files carry a governed identity from creation.Designed. Third-party packages and registries are outside what we show.Partial: designed or explainedOpen cited evidenceStage 4 Stage 9
ASI05 Unexpected Code Execution (RCE)Staged work is not authorized work: only accepted, authorized tasks reach a worker.Designed for authorization. Sandboxing and network isolation are outside our layer.Partial: designed or explainedOpen cited evidenceStage 2 Stage 6
ASI06 Memory & Context PoisoningThe context an agent works from is the oversight record and canonical records, carried with the task, not recalled from the agent's own memory.Recorded in part: the task contract and its records now travel in the agent's first message (2 Oct 2026).Recorded, in partOpen cited evidenceStage 3 Stage 4
ASI07 Insecure Inter-Agent CommunicationAgents talk through a governed relay; questions and answers stay tied to the same request on the Response Table.Recorded in part: one desk carried one request to three specialist agents, with a person arming each request (2 Oct 2026). Detecting unsanctioned side channels is not shown.Recorded, in partOpen cited evidenceStage 6 Stage 8
ASI08 Cascading FailuresThe truth gate stops a chain when a claim of completion is not backed by evidence; governance state cannot jump from healthy to hostile without passing intermediate states.Recorded: the gate caught a false "done" and stopped the chain, then exposed its own gap when it counted an honest "no data" reply as completed (1 Oct 2026, two local runs); the gap was fixed on 2 Oct 2026.RecordedOpen cited evidenceStage 7
ASI09 Human-Agent Trust ExploitationA reply is not completion. The person decides on the Response Table from the record, and every page labels what kind of claim it makes.Recorded: completion is now checked against criteria; a claim with no evidence fails (2 Oct 2026).RecordedOpen cited evidenceStage 7 Stage 8
ASI10 Rogue AgentsEach request is authorized by a person, results are checked against the record, and credentials lose validity as governance state degrades.Designed. Containment of an agent that has escaped its environment is outside our layer.Partial: designed or explainedOpen cited evidenceStage 2 Stage 7

Outside our layer: Sandboxing and network egress control; Protection of binaries against reverse engineering; Containment of agents that have escaped their environment. Qensai governs authority and evidence; it works alongside containment, not in place of it.

Reading the labels

Every page says what kind of claim it makes. Read the limits each page states.

Recorded proof Labelled per stage Illustration Explanation Explanation · sample data Operator diagram Live read view Machine-readable Analysis of public sources sign-in
Optional advanced review · recorded partial result

Put the governance claims under pressure.

The Governed Escape Test is for teams actively comparing Qensai with an existing or planned agent system. Inspect the declared success conditions, failure boundaries and machine-readable proof contract, then give the reviewed evaluation prompt to an agent you control. It comes after the story, proof and map—not in place of them.

← Back to the storyText guideMachine-readable routesOperational ComponentsQensaiContact