Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for AI agent developers, automation builders, ops teams, and power users who need a single agent that runs locally and reaches them on whichever channel they already use (Telegram, Slack, Discord, WhatsApp, WeChat, Lark/Feishu, Email, Webhook), with durable memory that survives across sessions, sandboxed code execution, and a SkillForge skill browser that ships with reusable skills for comm
Who should use it
Who should skip it
Hold off on EverMind-AI/Raven for mission-critical workflows without a containment strategy, explicit approvals, and a hands-on security review.
About this signal
EverMind-AI/Raven is tracked by RepoRadar as a self-improving agent harness in the AI Agents section. It was first seen on 2026-07-03 and last updated on 2026-07-03. The current verdict is 'try now' with a Silver tier and moderate setup difficulty. EverMind-AI/Raven leads on workflow potential (9.4) and open-source/build quality (8.4); its lowest signal is maturity (5.8), so factor that in before investing setup time. This page summarizes the evidence RepoRadar has captured from captured source metadata. The score, tier, risk label, and verdict on this page are never influenced by sponsorship, ads, or tips — they reflect only the usefulness, popularity, novelty, momentum, maturity, and evidence signals described in the RepoRadar methodology.
How this item is evaluated
RepoRadar assigned EverMind-AI/Raven a composite score of 7.9 out of 10, placing it in the Silver tier. This score combines weighted sub-signals: usefulness (35%), novelty (18%), momentum (14%), maturity (10%), open-source/build quality (7%), evidence quality (6%), workflow potential (6%), and setup ease (4%). Popularity is tracked separately at 1.0 and never affects the composite score or tier. The risk label of 'medium' reflects inherent user-impacting hazards, not generic novelty. Items with no risk flag may still require normal code review before production use.
Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.
Risk explanation
The 12 messaging-gateway adapters include WhatsApp, WeChat, and QQ — these are personal-messaging platforms whose Terms of Service restrict automated access. The bundled WhatsApp bridge and Weixin CLI id are present in the repo, but operating a personal-messaging bot against WhatsApp / WeChat / QQ in production may violate the platform's ToS. The Telegram / Slack / Discord / Matrix / Email / Webhook / Lark / DingTalk / WeCom / QQ-work channels are the lowest-risk starting set; review the ToS for the personal-messaging adapters before enabling them in production; Sandboxed code execution is the right default for an agent harness, but the sandbox boundary depends on the install context (Linux/macOS sandbox vs Windows native vs WSL2). Review the `raven onboard` sandbox configuration before exposing the agent to untrusted prompt sources, and pair with the `raven sentinel status` proactive-memory + scheduled-nudges view to audit what the agent scheduled to run unattended; The EverOS durable memory carries user-level + agent-level + world-level knowledge across sessions, which is the right primitive for self-improving workflows but means sensitive content (chat history, file paths, API keys referenced in commands) is persisted locally — review the `~/.raven/` and EverOS memory-store paths on shared / multi-user hosts and clear the memory store when decommissioning the agent.