CH-00
first builder loop
5 steps
Gate verifiedspine + electives. structured advancement.
Every team member runs the spine in sequence. Pilot gates use verified artifacts and rubric review. Electives go deeper by domain after the shared spine.
Required spineall team members, in sequence, pilot-gated
CH-01
terminal cockpit
6 steps
Gate verifiedCH-02
ai workflow
7 steps
In progressCH-03
git safety
5 steps
LockedCH-04
build useful machines
8 steps
LockedCH-05
builder proof project
6 steps
Capstoneevidence gates
Each pilot gate asks for an artifact reviewed against a rubric. Team leads can preview how submissions, revisions, and unlocks should work before a broader rollout.
Electivesunlocks after spine completion, choose by domain
css and design taste
Visual output, layout judgment, readable interfaces.
See detailsweb tool surfaces
Build tools that expose a UI, not just a prompt.
See detailsagents and evals
Multi-step tasks, evidence loops, output verification.
See detailsshipping projects
From local to deployed. Artifact visibility and handoff.
See detailstypescript builder logic
When code helps a tool. Just enough types to stay safe.
See detailscontext workbench
Wrangle the inputs. Templates, examples, retrieval.
See detailsai stack and model harness
Swap models without breaking a tool. Compare on real evals.
See detailsteam facilitation
For team leads. Rubrics, cohort cadence, gate reviews at scale.
Notify meEnterprise track · pilotorganizational best practices for AI rollout
shared context, governed agents, real workflows.
This track is for organizations deploying AI into mission-critical work, not individuals trying a better prompt. It teaches the operating habits between IT, security, operations, and business owners.
shared context systems
Source packets, decision logs, examples, retrieval boundaries, and team memory that survive beyond a chat thread.
access and egress maps
Classify read scope, write scope, external sends, destructive actions, approvals, receipts, and rollback paths.
agent observability
Run logs, traces, eval samples, escalation states, and dashboards for what agents did, changed, skipped, or need reviewed.
human-in-the-loop gates
Design review moments for customer impact, money, production writes, legal exposure, and sensitive data handling.
workflow change management
Turn a one-off AI demo into a repeatable operating process with owners, training, adoption metrics, and failure paths.
model upgrade operations
Re-test prompts, evals, cost routes, token budgets, and model assignments when capabilities or pricing change.
“Smart teams do not need more AI hype. They need a shared practice for turning messy work into tools they can trust.”
PromptDojo team curriculum
Ready to run your team through the spine?
Pilot cohorts start by arrangement. We help you set rubrics and pace before kickoff.