17 AGENTIC SDLC L400 FIELD GUIDE Download PPTX

STORY → OVERVIEW → BUILD → PROVE → DEPLOY

See the actual
policy engine work.

An animated, executable reference diagram. Its route, reviewer list and verdict come from the same JavaScript modules used by the CLI and tests.

Authenticity: this is our running reference implementation, not a reconstruction or screenshot of Levi Strauss's internal platform. Calculations are real; review records are explicitly synthetic. No AI agents execute here.

Follow one change through the system

LIVE REFERENCE IMPLEMENTATION
Executable evidence-lab flow Changed paths flow through risk routing, synthetic review fixtures, failure injection and structural validation. The final human-authorization boundary always remains closed. 01 / INPUTChanged pathsRepository-relative + exact SHA 02 / ROUTE5 / 8 / 11 reviewersActual route() calculation 03 / LAB REPORTSSynthetic fixturesNo AI agents executed 04 / MUTATEFailure injectionStale / missing / contradictory 05 / VALIDATEFail-closed structureActual validateEvidence() 06 / AUTHORITY: CLOSEDHuman release gateNot granted by this simulation ADVISORY EVIDENCE DOES NOT CONFER AUTHORITY

Real browser rendering · actual policy code · synthetic evidence · releaseAuthorized: false

Computed reviewer set

Expected SHA: ·

    Inspect the actual JSON output

    BUILD THE EXACT EXPERIENCE

    Code, commands and reproducibility

    01 / ENGINE

    Read the real code

    The page imports the unchanged policy and scenario modules copied by the build. No hardcoded “success” animation substitutes for a verdict.

    Open the policy source
    02 / UI

    Connect state to SVG

    Each animation step highlights an actual flow stage. The selected paths and injected failure drive the reviewer chips and JSON output.

    Open the animation source
    03 / PROOF

    Capture the running page

    Playwright interacts with the built site, asserts real outputs and saves unedited browser screenshots plus provenance.

    Follow the build walkthrough
    FROM A FRESH CLONE
    npm ci
npm test
npm run build
npx playwright install chromium
npm run browser:test
npm run capture
npm run build
npm run preview

    Open http://127.0.0.1:4173/workshop.html. Capture runs locally; it does not provision cloud services or invoke paid agents.

    ACTUAL CAPTURES, NOT MOCKUPS

    What the running implementation looks like

    These images are generated by the repository's browser-capture command. They show this reference site and its real validator output—not the source company's private system. Rebuild and recapture them yourself.

    Actual story-first overview in the running reference site.
    Actual story-first overview in the running reference site.
    Actual high-risk route: synthetic lab structure pass; authority remains closed.
    Actual high-risk route: synthetic lab structure pass; authority remains closed.
    Actual validator rejects reports after the expected head SHA changes.
    Actual validator rejects reports after the expected head SHA changes.
    Actual policy source loaded from lab/policy.mjs at site build time.
    Actual policy source loaded from lab/policy.mjs at site build time.

    Open the recorded six-stage animated GIF (actual browser frames; interactive controls are above).

    Captured 2026-11-10T00:05:29.381Z with Playwright 1.56.1 / Chromium (chrome channel, 156.0.8078.80). Unedited browser output. Capture provenance and SHA-256 fingerprints.

    Next: publish on GitHub Pages