pi-agent-dashboard / presentations
AGENTS.md349 starsChanged 23 days agoMITView on GitHub
Author-side deck sources. Each subdir = one deck3d presentation. Build with packages/deck3d/bin/deck3d build <dir>/deck.md -o <dir>/deck.html. Tune via inline <!-- deck3d: {...} --> overrides, never by editing deck.html.
In this repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- docker/AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- .github/codeql/AGENTS.mdAGENTS.md
- .github/codeql/extensions/AGENTS.mdAGENTS.md
- .github/verdaccio/AGENTS.mdAGENTS.md
- .github/workflows/AGENTS.mdAGENTS.md
- mockups/AGENTS.mdAGENTS.md
- public/AGENTS.mdAGENTS.md
- qa/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- spike/AGENTS.mdAGENTS.md
- tests/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
The file
Files
DOX — presentations
Author-side deck sources. Each subdir = one deck3d presentation. Build with
packages/deck3d/bin/deck3d build <dir>/deck.md -o <dir>/deck.html. Tune via
inline <!-- deck3d: {...} --> overrides, never by editing deck.html.
| File | Purpose | |
business-next-5-years/deck.md | Source outline. 23 slides, 25 min. Business change 2026-2031 for software/IT-services sales. 4 mermaid diagrams (3 flowchart, 1 sequence). Front-matter: dark/blackbelt/glass/high. Runs the front-matter blackbelt / the midnight deck override is GONE: midnight's low-contrast blue flattened the specular + reflection response the lighting rig exists for. See change: deck3d-cinematic-worlds. | |
business-next-5-years/deck.json | Derived IR. Regenerated by parse/build. Do not hand-edit below slides[]. A running deck3d serve re-parses deck.md and REWRITES this file — stop the server before editing it, or the edit is overwritten (kill by PORT: `lsof -ti :4321 | xargs kill; pkill -f "deck3d serve"misses the real process, whose cmdline isdeck3d/src/cli.ts serve`). See change: deck3d-cinematic-worlds. |
business-next-5-years/deck.html | Rendered offline deck. Generated artifact, NOT tracked in git (.gitignore) — rebuild with deck3d render deck.json -o deck.html. Committing it put CodeQL alerts (bundled three.js) and a Biome file-size warning into CI, and churned ~2.9 MB per rebuild. check: clean + style: 23/23 slides styled. See change: deck3d-cinematic-worlds. | |
business-next-5-years/ camera | Per-slide camera.distance 9-12 (was a flat 12-15). Fog is fixed at 14-36, so content parked past ~12 sat inside the ramp and washed out. Tuned by the loop "start every slide at 9, bump only the slides check reports a fit error for" / a blanket 9 overflows 9 of the 23. Lives in the deck.md INLINE overrides, NOT deck.json: markdown wins over the IR, so tuning written to deck.json alone is outranked and silently lost on the next parse. See change: deck3d-cinematic-worlds. | |
business-next-5-years/ (no fx/) | Deck carries NO local effects. All eight promoted into the corpus (deck3d/src/fx/) via fx promote; deck references them as plain corpus presets {id:"<name>"}, no sha256 pin. Renamed on promote: horizon-2031→horizon-gates, workforce-100→cohort-grid. Deck-specific 0.59 cohort split now passed as params.share, not baked in. See change: deck3d-cinematic-worlds. | |
business-next-5-years/speaker-notes.md | Per-slide timing table (25 min), Q&A prep, source list. | |
business-next-5-years/s*.png | Snapshot previews of styled slides (1, 6, 9, 15, 16, 20) after the Style pass. Regenerate with deck3d snapshot deck.html --slide n -o sn.png. See change: deck3d-cinematic-worlds. |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

