01 Designing how humans and agents work together

The agents act. The humans set the terms.

29 patterns now describe how that works in shipping products, from Claude Code to Codex and ChatGPT. Four have settled. Nine are picking up this month.

Agents & Humans counted 199 sightings of them in the last 30 days, recorded in daily field notes, and names the patterns that hold in essays.

FIG. 0 · THE STATE OF THE FIELD 29 PATTERNS · SIGHTINGS IN THE LAST 30 DAYS · AS OF OCT 25, 2026 0 10 20 30 40 HANDOFFSUNDERSTANDINGDECISIONSTRUSTCONTEXTMANY AGENTSSURFACES Observability 16 IN 30 DAYS · STEADY Approval 26 IN 30 DAYS · STEADY Review 14 IN 30 DAYS · PICKING UP Confidence Signaling 15 IN 30 DAYS · PICKING UP
Top insight · from 199 sightings across 29 patterns

Confidence Signaling doubled from 6 to 15 sightings in a month, the fastest climb in the library. Approval leads at 26 and has held steady. The 4 settled patterns carry 20 percent of everything seen this month.

4 of 29 patterns have settled, 9 are picking up, 199 sightings in 30 days FIG. 0 · THE STATE OF THE FIELD · OCT 25, 2026

solid = established halftone = emerging outline = still watching hairline = the 30 days before beacon = picking up One tower per pattern, height = confirmed sightings in published field notes over the last 30 days, grouped by family · n = 199 sightings across 29 patterns · snapshot 2026-10-25.

02

The Agent Workbench

INSTALL →

Run one agent from plan to review. Each panel is a component you can install, wired on the Interaction Contract and running against a simulated runtime in your browser. Nothing leaves the page.

THE WORKBENCH EXPLAINED: THE PEOPLE IT SERVES AND THE REASONS FOR ITS SHAPE →

1 Plan

Records agent wants to move 42 customer records from Sam Ortiz, who left, to Dana Reyes, then email each customer their new contact. Exclude records and see the fields that change.

Fields that change on each record
FieldNowAfter
OwnerSam OrtizDana Reyes
Renewal contactsam.ortiz@example.comdana.reyes@example.com

What the agent may touch

Object
Account
Write
ownerrenewal_contact
Read
namecityownerrenewal_contact
Denied
All other fields on Account, and all other objects.
Records in the plan

42 of 42 records included.

Proposed plan, version 1

  1. Read the 42 records owned by Sam Ortiz
  2. Set owner and renewal contact to Dana Reyes on 42 records
  3. Email each updated customer their new contact
  4. Write the change log

Built from

npx shadcn@latest add https://agentsandhumans.ai/api/registry/plan-preview.jsonPlanPreview sourcePlanPreview in Storybook
Interaction Contract log, 0 events

    03

    This week in the field

    ALL FIELD NOTES →
    OCT 23 Clipped via Vercel Blog Vercel gives its agent narrow, revocable access to production The agent stops at its permission boundary and asks, so the grant stays visible and revocable instead of expanding on its own. If you build on top of an existing agent, that stop behavior is yours to add. → AGENT IDENTITY OCT 23 Paper via arXiv Groups of AI agents copy each other into wrong answers High confidence from a group of agents can mean they all copied the first one. The review surface needs to show where agreement came from, or the human deciding on the output has nothing real to judge. → CONFIDENCE SIGNALING OCT 23 Paper via arXiv DoorDash ran 35,000 code changes through AI agents for $4.79 each The reviewer approved or rejected based on a summary the agents wrote, not on direct inspection of the work. The handoff summary is the interface, and that makes its accuracy the design problem. → AGENT-TO-AGENT OCT 23 Paper via arXiv Open index maps how much of a job agents can take over This index shifts the design question from whether an agent touches a task to how much of it a human can hand off. Reach for approval and review patterns before you let an agent run without a human checking the result. → APPROVAL OCT 23 Paper via arXiv Atlas gives designers a map of what a long agent run did The agent ran; the human arrived after. Atlas shows that the review screen is its own design problem, separate from the run itself or the progress view. Most products have not built it yet. → HUMAN HANDOFF OCT 23 Paper via arXiv ABRA gives each agent its own reusable screen panel The view is bound to the agent that made it. That one design move changes what a human can do: they can see who built the surface before they touch it. Generative-ui without that binding leaves the human with no footing. → AGENT IDENTITY

    04

    Essays

    ALL ESSAYS →

    SIGHTING · OCT 23

    DoorDash's reviewer approved 35,000 changes without reading them

    The summary an agent writes about its own work is now the thing a human judges. That handoff document is the design problem most teams have not named yet.

     → REVIEW

    SIGHTING · OCT 22

    30 audited platforms ask what an agent can do, not what it shares

    Approval screens ask what an agent may do, then stop. The question of what an agent may share goes unasked, and a new wave of research shows what breaks when that question has no answer.

     → PERMISSIONS

    SIGHTING · OCT 21

    Designers said yes to errors they did not look at

    A researcher watched reviewers look at an approval screen, click yes, and miss the error in front of them. That finding breaks the premise of a review step built on presence alone.

     → APPROVAL

    05

    The field, measured

    ALL INSTRUMENTS →

    Where the month moved · by family

    Understanding fell the furthest, from 46 to 27 sightings in 30 days.

    0 20 40 60 80 Decisions 72 +9 Trust 48 +8 Understanding 27 −19 Handoffs 18 −2 Context 15 −4 Many agents 14 +2 Surfaces 5 +1

    the 30 days before the last 30 days

    Table
    FamilyLast 30 days30 beforeChange
    Decisions7263+9
    Trust4840+8
    Understanding2746−19
    Handoffs1820−2
    Context1519−4
    Many agents1412+2
    Surfaces54+1

    Climbers and coolers · change in sightings

    14 patterns gained sightings and 11 lost ground against the month before.

    0 Confidence Signaling +9 Approval +5 Review +5 Agent Identity +5 Observability −13 Off-Brief Alert −8 Permissions −7

    gained · right of zero lost ground · left of zero

    Table
    PatternLast 30 days30 beforeChange
    Confidence Signaling156+9
    Approval2621+5
    Review149+5
    Agent Identity116+5
    Observability1629−13
    Off-Brief Alert210−8
    Permissions1522−7

    All three figures read the 2026-10-25 snapshot of pattern_drift_daily: confirmed sightings in published field notes, the last 30 days against the 30 before · n = 199 sightings this month, 204 the month before, across 29 patterns in 7 families.

    The pattern library A cabinet with a drawer per pattern family holding 29 patterns; the drawers with the 13 patterns being watched are pulled out.
    Approval is the most-sighted pattern of the last 90 days
    Approval
    47none prior
    Observability
    46none prior
    Permissions
    39none prior
    Review
    23none prior
    Confidence Signaling
    21none prior
    Context
    21up from 1
    Human Handoff
    21none prior
    Agent-to-Agent
    20none prior

    Published, annotated sightings attached to each pattern, trailing 90 days (Jul 28 – Oct 26). The 1px tick on each track is the prior 90 days (Apr 29 – Jul 28). n = 238.

    View as table
    PatternTrailing 90dPrior 90d
    Approval470
    Observability460
    Permissions390
    Review230
    Confidence Signaling210
    Context211
    Human Handoff210
    Agent-to-Agent200
    Where the momentum came from, week by week
    Approval8 this week
    Observability3 this week
    Permissions3 this week
    Review5 this week
    Confidence Signaling2 this week
    Context1 this week
    Human Handoff5 this week
    Agent-to-Agent5 this week

    Weekly counts of published, annotated sightings per pattern over the last 12 weeks (Aug 10 – Oct 26), every row on one scale; the dot is the latest week. Hover a week to read every pattern at once. n = 238.

    View as table
    PatternAug 10Aug 17Aug 24Aug 31Sep 7Sep 14Sep 21Sep 28Oct 5Oct 12Oct 19Oct 26
    Approval000832632878
    Observability000758925253
    Permissions100842732453
    Review000121522145
    Confidence Signaling000221143152
    Context000331603221
    Human Handoff010400430045
    Agent-to-Agent000231213125

    The field is still forming: only 4 of 29 patterns are established.

    established · 4 · stable across independent vendorsemerging · 12 · seen across products, shape settlingwatching · 13 · candidate, accumulating sightings

    Current lifecycle stage of every catalogued pattern · stages per the promotion bar on /methodology · n = 29.

    How long maturity takes
    ESTABLISHEDEMERGINGWATCHING0481216202428Agent-to-Agent· 2 mo2Review· 2 mo2Observability· 2 mo2Budget Scoping· 2 mo2Multi-Agent Roster· 2 mo2Ambient Status· 2 mo2Confidence Signaling· 2 mo2Permissions· 3 mo3Emergency Stop· 3 mo3Reach Preview· 3 mo3Off-Brief Alert· 3 mo3Context· 4 mo4Generative UI· 4 mo4Background Execution· 4 mo4Delegation· 5 mo5Approval· 5 mo5Eval Run Setup· 5 mo5Release Decision· 5 mo5Planning· 6 mo6Memory· 6 mo6Progress· 7 mo7Human Handoff· 7 mo7Interruption· 8 mo8Recovery· 8 mo8Agent Identity· 9 mo9Case Drilldown· 26 mo26Production Drift· 26 mo26Eval Set Authoring· 28 mo28Comparison View· 28 mo28

    MONTHS SINCE FIRST SIGHTING

    First sighting = earliest observed date attached to the pattern · promotions from the pattern changelog · n = 29 patterns.

    View as table
    PatternLaneFirst sightedMonths since
    Agent-to-AgentEMERGINGAug 20262
    ReviewEMERGINGAug 20262
    ObservabilityEMERGINGAug 20262
    Budget ScopingWATCHINGAug 20262
    Multi-Agent RosterWATCHINGAug 20262
    Ambient StatusWATCHINGAug 20262
    Confidence SignalingWATCHINGAug 20262
    PermissionsEMERGINGAug 20263
    Emergency StopWATCHINGJul 20263
    Reach PreviewWATCHINGJul 20263
    Off-Brief AlertWATCHINGJul 20263
    ContextEMERGINGJun 20264
    Generative UIEMERGINGJul 20264
    Background ExecutionEMERGINGJun 20264
    DelegationESTABLISHEDMay 20265
    ApprovalESTABLISHEDJun 20265
    Eval Run SetupWATCHINGJun 20265
    Release DecisionWATCHINGMay 20265
    PlanningESTABLISHEDApr 20266
    MemoryEMERGINGMay 20266
    ProgressESTABLISHEDApr 20267
    Human HandoffEMERGINGMar 20267
    InterruptionEMERGINGMar 20268
    RecoveryEMERGINGFeb 20268
    Agent IdentityEMERGINGFeb 20269
    Case DrilldownWATCHINGAug 202426
    Production DriftWATCHINGAug 202426
    Eval Set AuthoringWATCHINGJul 202428
    Comparison ViewWATCHINGJul 202428

    06

    The pattern index

    ALL PATTERNS →
    Approval ● established Deciding, with enough in view, before an agent does a thing that is hard to undo. npx shadcn add @agentsandhumans/approval-gate Planning ● established Agreeing on the shape of the work before the agent does it. npx shadcn add @agentsandhumans/plan-preview Progress ● established Knowing what the agent is doing, how far along it is, and whether to step in. npx shadcn add @agentsandhumans/progress-trace Interruption ◐ emerging Changing an agent's course while it runs, without starting over. npx shadcn add @agentsandhumans/interruption Recovery ◐ emerging Putting things back, or taking over, after an agent makes a mistake. npx shadcn add @agentsandhumans/recovery Delegation ● established Handing an agent a piece of work so it understands the job the way you do. Context ◐ emerging Shaping what the agent knows before it starts, and seeing where that knowledge came from. Memory ◐ emerging Knowing what the agent remembers about you, and being able to correct it.

    ● ESTABLISHED — A STABLE CONVENTION ACROSS INDEPENDENT VENDORS · ◐ EMERGING — SEEN ACROSS PRODUCTS, SHAPE STILL SETTLING · ◌ CONTESTED — VENDORS ACTIVELY DISAGREE

    07

    WEEKLY · FREE
    FIVE-MINUTE READ
    NO ADS · NO NOISE

    The important shift isn't that AI can generate things. It's that software can increasingly act.