athena-jev
ATHENA gives coding agents reflexes - an open-source cognitive control layer powered by TypeSafe Jev.
How it uses Jev
ATHENA is an independent open-source cognitive control layer for coding agents, powered by TypeSafe Jev. It adds native OpenCode observability while keeping the agent in control by default.
Highlights
- Cognitive Control: AEGIS (safety), METIS (progress), NIKE (evidence), and EPISTEMICS (uncertainty) compute GO , DELIBERATE , VERIFY , and BLOCK .
- Passive by default: enforcementMode: observe records decisions such as WOULD BLOCK without interrupting OpenCode. Switch to enforce only when desired.
- Tool Router: predicts the smallest useful capability set. It defaults to observe , so all host tools remain visible.
- Native OpenCode V2 TUI: a compact white-and-blue sidebar; V1 remains supported with reduced UI capability.
- Languages
- TypeScript 63% · JavaScript 35% · Python 1% · PowerShell 1%
- License
- MIT
- Latest release
- ATHENA v0.1.0
- Last activity
- Sep 2026
- Added
- 2026-09-24
- Stars at snapshot
- 1
- Status
- Community listing
Source: community catalog. Details and metrics reflect the published community snapshot and are not a performance endorsement. Documentation details extracted from GitHub · the README · the project website.