First 30 Minutes
First 30 Minutes
Section titled “First 30 Minutes”The canonical evaluator path for understanding, inspecting, running, and verifying Topogram from this repo.
Status: current Audience: evaluators, technical leads, technical buyers, and agents doing a first pass Use when: you want the shortest useful route through the current repo before choosing a deeper proof or implementation task.
Run commands from the Topogram repo root. This path is intentionally narrow: it helps you judge whether Topogram gives humans and agents a smaller, verified operating surface than rediscovering the repo from scratch.
1. Understand
Section titled “1. Understand”Read these first:
- What Topogram Is
- Topogram Model
- Agent First Run, if you are evaluating agent work
The claim to evaluate is not “Topogram reads every repo perfectly.” The useful claim is: a reviewed app map can preserve product intent, ownership, contracts, proof, and focused context so humans and agents make changes from bounded evidence instead of broad guessing.
2. Inspect
Section titled “2. Inspect”Install dependencies if this checkout has not been prepared:
npm installThen ask the CLI for the current adoption state and agent briefing:
node ./engine/src/cli.js onboard . --jsonnode ./engine/src/cli.js agent brief . --jsonGood output:
onboardreturns anonboarding_planwith stages for init, check, audit-bundle, generate, and verify.agent briefreturnsread_order,edit_boundaries,first_commands, workflows, and SDLC policy warnings with portable paths.- Neither command writes artifacts, generates app output, or runs project verification unless you pass an explicit write/generate/verify flag.
3. Run
Section titled “3. Run”Validate the app map and inspect one focused packet:
node ./engine/src/cli.js check . --jsonnode ./engine/src/cli.js query slice ./topo --journey journey_greenfield_start_from_template --detail compact --format markdownGood output:
checkreportsok: truefor the current workspace.- The slice has a frame, read order, work items, proof plan, related records, and next commands for one bounded focus.
- The slice should be useful without requiring a reader to inspect every
topo/**/*.tgfile.
For a human-readable view of the same packet, render the static HTML cockpit:
node ./engine/src/cli.js query slice ./topo --journey journey_greenfield_start_from_template --detail compact --format htmlFor a rough context-size comparison, run:
node ./engine/src/cli.js query context-savings ./topo --journey journey_greenfield_start_from_template --detail compact --format markdowncontext-savings uses approximate local token estimates. Treat it as an
attention-budget signal, not exact model-token accounting.
4. Verify
Section titled “4. Verify”Run the smallest docs/repo proof for this evaluator pass:
npm run docs:checkGood output:
- documented command surfaces still match the current CLI;
llms.txtand the generated retrieval bundle stay in sync;- public docs keep stable headings, audience metadata, and current links.
When you want a portable evidence bundle instead of terminal output, write one:
node ./engine/src/cli.js onboard . --write --out-dir ./artifactsThe bundle should contain a manifest, check summary, agent brief, SDLC reports, context inventory, source excerpts, and portable paths.
Deeper Routes
Section titled “Deeper Routes”- Beta Demo Path: which public proof to run first.
- Proof Walkthrough: full proof catalog and how to read checkpoints.
- Brownfield Extract/Adopt: existing app discovery.
- Greenfield Generate: starting from authored Topogram or a template.
- Agent First Run: task-scoped agent workflow.
- CLI Reference: full current command map.