Practice that runs, not slides that scroll.
Every lab environment behind LockedIn Labs training is executable: a learner works a scenario and produces an artifact, and a reviewer scores it against a rubric published before submission. This page lists every environment in the catalog, what it practices, and whether it closes a node on the capability graph.
Every lab environment behind a seat, in one table.
Sorted by track, then level, then title. Counts derived from this repository.
| Environment | Track and module | Level | Minutes | Kind | Purpose | Closes a node? | Deliverable | Rubric dimensions |
|---|---|---|---|---|---|---|---|---|
| Evaluation Regression Lab | Agentic Workflows & AI Agents · Evaluating agent behavior | L3 | 55 | Evaluation regression challenge | Challenge | Practice only | Signed evaluation release-decision receipt | 4 |
| Hold or ship the support agent | Agentic Workflows & AI Agents · Evaluating agent behavior | L3 | 35 | Memo workbench | Curriculum | Yes | Release decision memo | 4 |
| Agentic Release Challenge | Agentic Workflows & AI Agents · Design an agent fleet | L4 | 60 | Agentic release challenge | Challenge | Practice only | Authorized release evidence receipt | 4 |
| Stop a self-certifying agent fleet | Agentic Workflows & AI Agents · Design an agent fleet | L4 | 90 | Memo workbench | Curriculum | Yes | Fleet release-control brief | 4 |
| Gate the agent-written migration PR | AI Native Developers · Reviewing machine code | L3 | 35 | Memo workbench | Curriculum | Yes | Review verdict memo | 4 |
| Recover the enterprise copilot pilot | AI Native Developers · Ship an agent to production | L3 | 90 | Memo workbench | Interview | Practice only | Technical recovery and production rollout plan | 4 |
| Run the forward-deployed field loop | AI Native Developers · Ship an agent to production | L3 | 75 | FDE delivery capstone | Challenge | Practice only | Versioned FDE capstone evidence packet | 6 |
| Gate a networked shell agent | AI Native Developers · Ship an agent to production | L4 | 90 | Memo workbench | Fieldcraft | Practice only | Containment and release authorization brief | 4 |
| Route a model swap by risk | Executives & Leaders · Where AI actually pays | L3 | 35 | Memo workbench | Curriculum | Yes | Model routing decision | 4 |
| Enterprise Data Reconciliation Lab | Forward-Deployed Engineering · Land inside enterprise systems | L4 | 65 | Enterprise data reconciliation | Challenge | Practice only | Reconciliation command rehearsal | 5 |
| FDE Production Project Review | Forward-Deployed Engineering · Operate adoption, rollout, and incidents | L4 | 90 | Production commission | Challenge | Practice only | Exact-commit project registration | 10 |
| Production Operations Defense | Forward-Deployed Engineering · Operate adoption, rollout, and incidents | L4 | 75 | Memo workbench | Challenge | Practice only | Production evidence defense brief | 8 |
| Contain an unregistered agent | AI Native SDLC & Control Plane · Authorization and the evidence chain | L4 | 40 | Memo workbench | Curriculum | Yes | Incident and authorization brief | 4 |
Practice-only environments never close a capability graph node. Node-closing environments require an artifact reviewed against the published rubric.
Four rooms, four field moments.
Each room rehearses a field moment with a scripted counterpart and produces a saved brief or after-action record. All four sit inside the authenticated workspace.
Rehearses a first client discovery call against a scripted counterpart, then produces a saved discovery brief inside the authenticated workspace.
Rehearses recovering a broken integration mid-engagement against a scripted counterpart, then produces a saved incident brief inside the authenticated workspace.
Rehearses commanding a staged rollout decision against a scripted counterpart, then produces a saved rollout brief inside the authenticated workspace.
Rehearses transferring ownership of a live system to a client team against a scripted counterpart, then produces a saved after-action record inside the authenticated workspace.
The rubric is published before you submit.
Three rules govern every node-closing environment.
The rubric renders before submission.
A learner sees exactly what a reviewer will score before writing the first line of the deliverable.
A failed review leaves the node in progress.
Resubmission is the designed path, not an exception.
An unchanged artifact is refused for re-scoring.
A learner cannot resubmit the same text and ask for a second verdict.
See the practice before you commit a program.
We walk your team through the exact lab environments and practice rooms behind the seats a program would staff.