List view
Three-tier object permissions (read / read-write / admin) on skills + skillsets, typed grants, and owner-to-owner ownership transfer in the Danger Zone. Non-disruptive: all existing shares migrate to read-only.
Due by July 6, 2026•5/5 issues closedAPI-first natural-language assistant over Ornn's own docs/positioning + the skill catalog; ornn-web widget is the first consumer (future: MCP tool). v1: authenticated-only, pure Q&A (no agentic tools), knowledge = CI-distilled curated KB + per-request scoped skill retrieval. Due date is a placeholder, adjust as needed.
Due by July 21, 2026•2/3 issues closedLayer 1: skill→skill prerequisites in metadata + closure resolver. Layer 2: typed, versioned skillsets (revive #916 with kinds, e.g. consensus-supported). Agent-side consensus consumes these; Ornn packages + delivers, does not execute (see #962).
Due by July 8, 2026•14/15 issues closedAuth/authz + security hardening surfaced by the /auto security-review audit (run 20260604T053431Z-72549). Crown jewels: NyxID identity-token verification (P0) and playground private-skill BOLA (P1).
Overdue by 2 day(s)•Due by June 18, 2026•18/21 issues closedSocial engagement layer on the skill registry: in-product skill ratings, text reviews, likes/upvotes, per-skill discussion threads, author reputation/profile, activity feed. Replaces the abandoned marketplace direction (#38 billing / #39 curated archived here as historical). Activates after launch when there are real external users. Outcome: skill discovery driven by community signal, not just admin curation.
No due date•2/2 issues closedSkills compose into workflows. Skill chaining (declarative pipelines), webhook + event system for skill-lifecycle events, templates & starter kits for common patterns. Deferred behind QA + Code Review hardening — comes after the launch gate. Outcome: a skill is no longer a leaf; it is a composable building block.
No due date•0/6 issues closedManual functional QA against ornn.chrono-ai.fun before public launch. Twelve module-level test trackers (login+permissions, skill creation+GitHub import, format validation, playground+sandbox, security checks, registry+detail, owner skill mgmt, admin dashboard+quota+codes, CLI+agent+NyxID proxy, notifications, admin skill mgmt+config, static pages). Explicit go-live exit criteria. Discovered bugs become standalone issues filed against the feature milestone they affect. Outcome: green checklist confirming every user-facing flow works end-to-end.
No due date•20/21 issues closed