Skip to content

ROADMAP.md

Latest commit

 

History

History
135 lines (89 loc) · 8.87 KB

ROADMAP.md

File metadata and controls

135 lines (89 loc) · 8.87 KB

EvoNexus Roadmap

Unofficial toolkit for Claude Code — AI-powered business operating system.

This roadmap is updated regularly. Want to vote or suggest? Open a discussion or create an issue.

Legend

Symbol Meaning
[ ] Not started
[x] Done
⚠️ Breaking change
🔥 High priority
💡 Needs design discussion first

v0.4 — Foundation & Stability

Fix, secure, and improve what already exists before growing.

Skills

  • 🔥 Evolution product skills — int-evolution-api (33 commands), int-evolution-go (24 commands), int-evo-crm (48 commands) for managing instances, messages, contacts, conversations, pipelines via REST API.
  • Version indicator & update alerts — show current version in dashboard sidebar, alert when new GitHub releases are available.

Developer Experience

  • 🔥 CLI installer — npx @evoapi/evo-nexus — clones repo, installs deps, runs interactive setup wizard.
  • Full Docker install — docker compose up dashboard with multi-stage Dockerfile + GitHub Actions CI pushing to GHCR.
  • Update checker — dashboard checks GitHub releases and shows upgrade notification.
  • settings.json — project-level permissions (allow/deny), hooks configuration, thinking mode enabled.
  • CLAUDE.md split — reduced from 263 to 128 lines. Detailed config moved to .claude/rules/ (agents, integrations, routines, skills).
  • Inner-loop commands — /status (workspace status) and /review (recent changes + next actions).

Dashboard UX

  • Sidebar reorganization — 5 collapsible groups (Main, Operations, Data, System, Admin) with localStorage persistence.
  • Active agent visualization — Claude Code hooks track agent launches via PreToolUse events, writing to agent-status.json. Dashboard polls /api/agents/active and shows "RUNNING" badges with pulse animation on agent cards and overview.
  • Agents page redesign — unique icons and accent colors per agent, status dots, slash command badges, memory count pills, hover glow effects.
  • Overview page redesign — stat cards with icons and trend indicators, active agents bar, quick actions row, improved reports and routines tables.

Agent Generalization

  • Agent prompts generalized — all 9 agent prompts cleaned of hardcoded personal references. User-specific context preserved in _improvements.md per agent memory folder.

v0.5 — Extensibility & Ecosystem

Make EvoNexus composable and self-extending.

Agent System

  • 🔥 Generalize existing agents — all 9 agent prompts generalized. User-specific context preserved in _improvements.md per agent memory folder. Adapter patterns documented as future work.
  • 🔥 New business agents — expand functional coverage:
    • Marketing Agent — orchestrate existing mkt-* skills, attribution, budget, full funnel
    • HR / People Agent — onboarding, 1:1s, performance reviews, hiring pipeline
    • Customer Success Agent — health score, churn prediction, NPS/CSAT, client onboarding
    • Legal / Compliance Agent — contracts, renewals, GDPR/LGPD, compliance checklists
    • Product Agent — discovery, feature prioritization (RICE/ICE), PLG metrics, feedback loop
    • Data / BI Agent — cross-area consolidated dashboard, unified KPIs, alerts, trend analysis
  • Custom agents — custom- prefix, gitignored, auto-discovered by dashboard (core/custom badges), create-agent skill, create-command skill
  • Help agent (Oracle) — /oracle answers questions about the workspace by reading the actual docs. No RAG needed — reads docs/llms-full.txt and source files directly

Routines & Scheduling

  • 🔥 Trigger registry — define and manage named triggers (webhook, cron, event-based) that invoke skills or routines
  • Non-recurrent scheduled actions — one-off scheduled tasks (e.g., "post this on LinkedIn Friday at 10am") without creating a full routine
  • Systematic routines — pure Python routines via run_script() — no AI, no tokens, no cost. create-routine skill generates the code

Integrations

  • Complete Obsidian integration — finish obs-* skills: bidirectional sync, canvas, bases, CLI

Import / Export

  • Backup system — export workspace state as ZIP (agents, skills, routines, memory, config); import to restore. Support local, git, and cloud bucket targets.

v0.12 — Engineering Layer

Add a complete software development team alongside the business agents.

Engineering Layer (delivered)

  • 🔥 19 engineering agents — complete dev team derived from oh-my-claudecode (MIT, Yeachan Heo). Reasoning tier (opus): apex-architect, echo-analyst, compass-planner, raven-critic, lens-reviewer, zen-simplifier, vault-security. Execution tier (sonnet): bolt-executor, hawk-debugger, grid-tester, probe-qa, oath-verifier, trail-tracer, flow-git, scroll-docs, canvas-designer, prism-scientist. Speed tier (haiku): scout-explorer, quill-writer.
  • 🔥 25 dev-* skills — Tier 1 orchestration (15): dev-autopilot, dev-plan, dev-ralplan, dev-deep-interview, dev-deep-dive, dev-external-context, dev-trace, dev-verify, dev-ultraqa, dev-visual-verdict, dev-ai-slop-cleaner, dev-sciomc, dev-team, dev-ccg, dev-ralph. Tier 2 setup (5): dev-mcp-setup, dev-deepinit, dev-project-session-manager, dev-configure-notifications, dev-release. Tier 3 utilities (5): dev-cancel, dev-remember, dev-ask, dev-learner, dev-skillify.
  • 15 dev templates — .claude/templates/dev-*.md for each agent's primary output (architecture-decision, work-plan, code-review, bug-report, verification-report, deep-interview-spec, security-audit, test-strategy, trace-report, explore-report, design-spec, analysis-report, research-brief, critique, simplification-report).
  • workspace/development/ folder — 7 subfolders (architecture, plans, specs, reviews, debug, verifications, research) for engineering layer artifacts. Distinct from workspace/projects/ (active git repos).
  • Two-layer architecture documented — .claude/rules/agents.md, CLAUDE.md, docs/agents/overview.md, docs/agents/engineering-layer.md, docs/architecture.md, docs/introduction.md, site Home.tsx.
  • Open source attribution — NOTICE.md at repo root with full MIT license, version pinned (v4.11.4), modifications listed. Credits in README.md.
  • Pattern compliance — all 19 engineering agents follow the EvoNexus standard: rich frontmatter (Examples + commentary), Workspace Context, Shared Knowledge Base, Working Folder, Identity, Anti-patterns, Domain, How You Work, Skills You Can Use, Handoffs, Output Format, Continuity. Verified by @lens-reviewer (3 fixes applied: oath-verifier disallowedTools, raven-critic and trail-tracer Skills section).

Cross-layer pipelines (now possible)

  • End-to-end implementation — dev-autopilot orchestrates spec → plan → code → QA → validation across multiple engineering agents.
  • High-stakes consensus planning — dev-ralplan runs Planner/Architect/Critic consensus loop with RALPLAN-DR structured deliberation.
  • Bug investigation — @trail-tracer (multi-hypothesis) → @hawk-debugger (root cause + minimal fix) → @oath-verifier (regression check).
  • Pre-merge gate — @lens-reviewer (code quality) → @vault-security (OWASP audit) → dev-ultraqa (build/test/fix loop) → @oath-verifier (acceptance criteria).

v1.0 — Community & Growth

Community adoption, discoverability, and self-sustaining ecosystem.

Community & Docs

  • 🔥 Public roadmap — this file. Community input welcome via discussions.
  • Telegram & Discord channels — activate community channels, document in README and docs site.
  • In-app tutorials — contextual tutorials surfaced inside the dashboard, not just external docs.
  • Resume Claude sessions in chat — list active/resumable Claude sessions in dashboard chat with --resume support.

Development

  • Testing framework — define and implement test strategy for skills, routines, and agent behaviors; prevent regressions.

Contributing

Want to help? Pick any [ ] item and:

  1. Check open issues
  2. Read CONTRIBUTING.md
  3. For 💡 items, open a discussion first — design is still open

Last updated: 2026-04-10 — Evolution Foundation