Build video: Govern Roles with Role Groups and PIM

[dahans-msft2]

Build video: Govern Roles with Role Groups and PIM

A ~4-minute unit companion video for Provision and govern identities in Microsoft Entra.

Prep is approved and committed on this branch (rebuilt on the fixed pipeline, main @57aa4d04): script.md, design-plan.md, narration.wav (en-US-Vance, Dragon HD), transcript.json.

Build task (in learn/output/Govern-Roles-With-Role-Groups-And-PIM/):

1. Fill every chrome __FILL__ from script.md.
2. Author the body scenes per design-plan.md, inserting them into scenes.json's body_slot.
3. Anchor every beat + in-scene cue to transcript.json word times.
4. Assemble + pass gates from the project dir: assemble_scenes -> check_subcomps -> check_placeholders -> check_initial_state -> check_cue_anchors -> lint -> check.

Do not render, change narration or palette, edit learn/tools/**, or touch anything outside learn/output/Govern-Roles-With-Role-Groups-And-PIM/. Follow .github/copilot-instructions.md and .github/agents/hyperframes-builder.agent.md.

To start: comment @copilot on this PR with your chosen model. When the session ends and the PR is marked ready, the render auto-fires (or add the render label / comment /render).

[dahans-msft2]

@copilot

[Copilot]

@copilot

Build complete — composition assembled and every gate green (6c8dd0a).

What was built (all in learn/output/Govern-Roles-With-Role-Groups-And-PIM/):

• Chrome __FILL__ filled from script.md — bumper kicker, title lockup, 3 objective steps, recap payoff, CTA.
• 9 body scenes authored per design-plan.md and inserted into the body_slot: 04-roles-today + 07-assignment-types (list-select), 05-role-groups / 10-lockout-risk / 12-custom-role (callout-note), 08-activation (list-steps), 09-approver-puzzle (console-status), 11-durations (list-specs), and 06-always-on authored custom on _foundation.css. 10-lockout-risk is the one dark-field scene.
• anchors.json rewritten (82 anchors) and anchors.js regenerated with word_anchors.py — every beat start and every in-scene cue position is a real spoken word time from transcript.json, converted to scene-relative. No words-per-second estimates anywhere.
• scenes.json rewritten: 14 scenes, 279.4s + the 10.667s AI end card = 290.067s composition; narration mounted at the 3.0s lead-in.

Gates (from the project dir, CLI 0.7.77):

gate	result
assemble_scenes.py	14 scenes, 12 seams → index.html
check_subcomps.py	MOUNT-CONTRACT OK — ids triangulate, track 1 tiles cleanly
check_placeholders.py	no unfilled placeholders
check_initial_state.py	14/14 clean
check_cue_anchors.py	14/14 fully anchored, 100% coverage
lint	0 errors, 0 warnings
check	Runtime 0 · Layout 0 issues · Motion 0 · Contrast 18/18 WCAG AA

One escalation for you (shared tool, not patched here): learn/tools/assemble_scenes.py:194 calls ff = font_faces(fonts) without forwarding base, so the "skip the @font-face when the WOFF2 is absent" guard added in 910bdf8 never fires — the cloud gate then 404s on fonts/segoeui-*.woff2, which are git-ignored by design. I worked around it inside this project only (scenes.json → "fonts": [], with the two original faces preserved under _fonts_local and a _fonts_note); the kit foundation already declares font-family: "Segoe UI", system-ui, sans-serif, so the local render still picks up real Segoe. The one-line tool fix is yours to make since learn/tools/** is read-only for me.

Not rendered, narration and palette untouched, and the diff is confined to this project folder.

[github-actions]

🎬 Render complete — download the MP4 + captions + thumbnail from this run's artifacts (video-31243452722): https://github.com/dahans-msft2/hyperframes/actions/runs/31243452722#artifacts

Review it, then merge to videos to archive.

[dahans-msft2]

Closing: these were model-comparison tests, and the branch->videos-archive model is being replaced (the scope gate false-fails on base drift). Branch retained for reference.