The product is built by AI agent sessions playing defined roles under a written rulebook, with a human (Dinis Cruz) making the decisions only a human can. Every role's outputs are committed; the queue below is live.
An agentic team without structure produces the same mess as a human team without structure — faster. The structure here is simple: each role has a written ROLE file (mandate, boundaries, outputs), a folder for its reviews, and the rulebook binds them all to the reality discipline. When a role has produced nothing, its empty folder says so — the structure is honest about its own exercise.
| Role | Owns |
|---|---|
| Conductor | Routing every task across roles; priorities |
| Architect | The load-bearing decisions: no backend, key-as-billing, privacy modes |
| Dev | The product front end, built from the existing SGraph tools |
| Designer | The one-motion experience — form, function, intention |
| DevOps | The two-branch CI, auto-tagging, both deploy estates |
| Librarian | Reality vs proposed; the library; the master index |
| QA | Test strategy; the audio-format matrix is test surface #1 |
| AppSec | Key handling: caps, rotation, revocation, the seeded-key abuse model |
| DPO | The privacy-mode claims — routed / restricted / browser-local must be true |
| Ambassador | Growth and positioning: sell the availability gap, honestly |
| Journalist | The site's Updates section; the per-version changelog, with the Librarian |
Open and blocked issues as of this deploy. Each is a markdown file in the repo — click through for the full brief, status history and outcome criteria.
| # | Title | State | Priority | Created |
|---|---|---|---|---|
| loading issues.json… | ||||
The full queue including everything done is on the hub page.