ADHDev Model Card
这是可索引项目报告证据页：它保留 ADHDev Model Card 的源材料入口、机制线索和限制提醒；正文仍需 reader/editor 与 academic public-copy review 后才能当作最终结论引用。
ADHDev Model Card
| Field | Value |
|---|---|
| Repository | vilmire/adhdev |
| Category | Coding-Agent Control Plane |
| Stars / forks snapshot | 33 / 8 |
| Language | TypeScript |
| License | AGPL-3.0 |
| Raw capture | raw-github/vilmire_adhdev.md |
| Updated by | hourly public metadata update, 2026-05-26 05:44 +0800 |
1. Role in Self Evolve
ADHDev is not a self-evolution algorithm, but it is relevant infrastructure for long-running coding agents: it gives users a local dashboard to monitor, interrupt, resume, and inspect multiple agent sessions.
2. Working Principle
coding-agent session -> local daemon/dashboard -> status, terminal, chat, approvals, history -> human-controlled continuation or interruption
3. Evidence Path
Web-observed GitHub evidence showed 33 stars, 8 forks, 1,127 commits, 243 tags, AGPL-3.0 licensing, and a TypeScript-dominant implementation. README states it supports Cursor, Google Antigravity, VS Code, Kiro, Codex, Claude Code, Hermes Agent, Codex CLI, and Claude Code CLI as tested or working integrations.
4. Teaching Use
Use this card to explain the operational control-plane layer: self-improving coding agents need observable sessions and approval states before they can safely run as background workers.
5. Limits
The repository was not run locally and no agent session was connected. Shell GitHub API DNS failed, so metadata is web-observed rather than API-verified.