Skip to content

Latest commit

 

History

History
64 lines (49 loc) · 4.54 KB

File metadata and controls

64 lines (49 loc) · 4.54 KB

Swarm Knowledge Protocol — Roadmap

This document tracks the strategic vision and concrete milestones for the SKP ecosystem. Short-term items are actionable; medium/long-term items are directional and open for community discussion.


Short-term (0–3 months) — Now

Area Item Priority Depends On
Quality scoring ✅ Phase A complete — 141 lessons scored, CI integrated P0 #210
A→C 闭环 fatal-guard tombstone → draft lesson pipeline, bench-core dynamic tasks P0 fatal-guard v0.2.0
Proof of Access Lightweight quota system — 5 free searches, contribution refill P1 profile.py
Governance Define trust tiers in machine-readable config (misaka-protocol.json) P0 #211
Contributor wall Fix time-decay formula — cap per-PR weight, remove contributor disappearance P0 #212
Trust system Formalize GitHub-verified vs mail-verified vs web-verified tiers P1 #213
Log Harvester CLI misaka harvest — stdin/bash-history → auto-generate SKP lesson ✅ Done #214
Ecosystem config Harvest tool interface declared in misaka-protocol.json P1 #215
Asset isolation Move docs/frontend assets to sidecar repository or /web P1 #216

Medium-term (3–9 months) — Next

Area Item Priority Depends On
Reputation system Reuse signal: count "👍 useful" clicks + cross-lesson references P0 #217
Reaction collection Frontend "👍 helpful" button → data/reactions.json (anonymous, git-backed) P1 #218
Cross-reference auto-detect CI scans lessons for [[lesson-id]] links, builds reference graph P2 Schema stable
Rank change notifications GitHub Issue auto-published when leaderboard #1 changes P1 #219
Personal rank alerts Notify individual contributors when they move up/down P1 #220
PR scoring Score PRs by complexity (files + lines + test coverage) P2 Quality gate stable
Seat belt mechanism Auto-cap contribution weight for very high-frequency contributors (sigmoid cap) P2 Leaderboard data >50 contributors

Long-term (9+ months) — Vision

Area Item Priority Depends On
Retrieval weighting Search-hit rate → contribution bonus P2 Frontend analytics
Hub federation Cross-repo lesson sync via Hub nodes P2 Hub mode stable
Plugin system External tools read/write lessons via SKP protocol P3 API surface stable
Leaderboard prize pool Top monthly contributor gets a badge / early-access feature P3 Community >100 nodes
CLI dashboard misaka dashboard — local terminal leaderboard + notifications P3 Harvester CLI stable
SKP SDK TypeScript SDK for browser-agent interactions P3 Hub API stable
DALN whitepaper Formal paper on Decentralized Autonomous Learning Networks P4 All above stable

Design Principles

  1. Offline-first. Everything must work in air-gapped environments. Features that require always-on servers get de-prioritized.
  2. Zero-dep core. The misakanet-core PyPI package stays dependency-free. All extras are optional.
  3. Git as the source of truth. No database. No daemon. No centralized API dependency.
  4. Agent-native. Configuration must be machine-readable first, human-readable second. Every human doc has a JSON equivalent.
  5. Anti-hype. We ship what we can verify. Every claimed capability must be demonstrable. Every limitation must be disclosed.

How to Contribute

  • Open a GitHub Issue with the roadmap label
  • Propose changes via PR with rationale
  • Vote on priorities by reacting to pinned roadmap Issues

README → · Limitations → · Protocol Config →