Skip to content

Add OIA Application Matrix (v0) — committee dashboard sibling page - #34

Draft
michaeloboyle wants to merge 4 commits into
mainfrom
feature/oia-application-matrix
Draft

Add OIA Application Matrix (v0) — committee dashboard sibling page#34
michaeloboyle wants to merge 4 commits into
mainfrom
feature/oia-application-matrix

Conversation

@michaeloboyle

Copy link
Copy Markdown
Collaborator

OIA Application Matrix (v0) — committee dashboard sibling page

Adds docs/oia-matrix.html, a self-contained interactive tool that maps community projects onto the Open Intelligence Architecture stack (10 layers × 6 cross-layer spans), linked from the dashboard header.

v0 draft — a strawman for committee review, not a measurement. Most placements are classified from repo descriptions (see the per-row evidence tier); some are unverified and may be wrong. A working tool to structure the OIA intake discussion, not an official Foundation ranking of any project.

Overview

overview

Click any row to expand its rationale inline (no scrolling)

drill-down

Each repo gets a gap-closing improvement prompt, grounded in the specs

gap prompt

What it does

  • Layer footprint — where each project sits (● centre of gravity, · presence), mirroring the OIA Provider Footprint.
  • Evidence tier — every placement is labelled code ✓ (checked in source), desc (description only), or desc ⚠ (code contradicts the description).
  • Gap prompts — per-repo, generated from actual gaps, with links to the OIA reader, digest, charter, constitution, RFC-001.
  • Client-side rescan — fetches the repo's live file tree and re-derives the footprint (heuristic auto tier); checks real files, not just names.
  • Styled to match the committee dashboard (cream / terracotta / Barlow Condensed / IBM Plex Mono).

Reviewer notes

  • Draft on purpose — held for committee review before it goes live on Pages. Merging publishes it.
  • It evaluates named repos (incl. ruvnet/*) via a solo v0 rubric; please red-line placements you disagree with. Independent re-tagging is the intended next step.
  • docs/oia-matrix.assets/ holds only the preview screenshots for this PR; can be stripped before merge.
  • Full write-up in PKM: Orgs/Agentics-Foundation/Analyses/2026-07-15-OIA-Application-Evaluation-Matrix.md.

🤖 Generated with Claude Code

A self-contained, committee-styled tool that maps community projects onto
the Open Intelligence Architecture stack (10 layers x 6 spans), with per-repo
drill-down, description-vs-code evidence tiers, gap-closing improvement prompts
(with spec links), and a client-side structural rescan.

Staged for committee review. The page carries a prominent v0 banner: it is a
strawman to structure the OIA intake discussion, not an official ranking.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Places the Intent-based Inter-agent Communication Protocol in the OIA stack:
centre of gravity at L2 (sovereign self-hosted mesh) and L5 (intent routing /
node discovery), presence at L3/L7/L9. Strongest spans: sovereignty, identity,
security. Evidence tier: desc (published spec v1.5 + live 2026-06-05 mesh probe,
not a source audit; mesh repo invite-only, which fails the openness gate).

First non-agenticsorg/non-ruvnet row in the matrix. Gaps capture the honest
live state: discovered nodes unreachable (relay_capable:false), mesh is 3 sub-1B
toy models, DID identity unshipped, A2A/ACP bindings unimplemented.

Screenshots (01 overview / 02 drilldown / 03 gap-prompt) regenerated to show IICP.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
@michaeloboyle

Copy link
Copy Markdown
Collaborator Author

Updated: added RobLe3/iicp.network (the IICP protocol) as the matrix's first external applicant — every prior row is an agenticsorg/* or ruvnet/* repo the committee already stewards.

OIA placement (evidence tier desc):

  • Centre of gravity: L2 Sovereign Infrastructure (decentralized, operator-self-hosted mesh) + L5 Inference & Retrieval (intent-URN routing / node discovery — its core function).
  • Presence: L3 (directory + .well-known/iicp-intents.json), L7 (dispatch beneath A2A/ACP, complementary to MCP), L9 (client SDKs).
  • Strongest spans: sovereignty, identity, security. (Identity is the set's weakest span, so IICP touching it is notable.)
  • Evidence note flags the honest state: classified from the published spec v1.5 + a live 2026-06-05 mesh probe, not a source audit. The mesh repo is invite-only — a client-side rescan will 404, which itself signals a failed openness gate.

Gap prompt captures the real live state: discovered nodes advertise score 0.91 but are unreachable to third parties (relay_capable:false); mesh was 3 sub-1B toy models on one host; Phase 3 DID identity unshipped; A2A/ACP bindings scoped but unimplemented (#202).

All three screenshots (overview / drill-down / gap-prompt) regenerated to show the IICP row. Placement is a v0 strawman for committee red-lining, consistent with the page disclaimer — and note IICP has not formally applied to intake; this is our structural classification of an external project.

michaeloboyle and others added 2 commits July 16, 2026 14:11
Adds a subtle ↗ affordance next to every repo name in the matrix (opens
github.com/<org>/<repo> in a new tab) and links the name in each row's
drill-down header. The row-name link uses stopPropagation so it does not
trigger the click-to-expand. Screenshots regenerated to show the links.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two Lion-ecosystem repos (Haiyang Li / khive.ai):
- lionagi: L7-centered orchestration framework, broad L2-L9 footprint,
  local-first sovereignty + resumable-run continuity. Evidence: desc.
- LNkernel: formally-verified capability-security microkernel (Rust+Lean),
  deep L8 continuity via machine-checked Lean4 proofs + L7 substrate.
  Evidence: code (source-audited; real proofs, unlike lean-agentic's claim).

Regenerated all 3 screenshots (drilldown features LNkernel).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants