Skip to content

DEVELOPER-EXPERIENCE.md

Latest commit

 

History

History
48 lines (36 loc) · 2.75 KB

DEVELOPER-EXPERIENCE.md

File metadata and controls

48 lines (36 loc) · 2.75 KB

Developer Experience

Back to README | See also: Architecture | Getting Started | Inference | Comparison

AIOS Developer Experience

hex presents system state through 4 progressive disclosure layers (ADR-2604131500), so developers get the right level of detail without information overload:

Layer Surface What You See
Pulse Statusline / glance One-line health: `hex: A+ 100/100
Brief hex brief Structured summary — architecture score, active swarms, recent completions
Console hex status / dashboard Full detail — every agent, task, inference call, boundary violation
Override hex steer / hex override Emergency controls — pause swarms, kill agents, force-escalate tiers

Trust delegation lets hex make decisions autonomously within bounds you set. Configure how much latitude agents get — from "ask me everything" to "handle T1/T2 autonomously, notify me on T3." The taste graph learns your preferences (formatting, naming conventions, error handling style) and applies them to generated code without explicit instructions.

hex trust show                  # Current delegation level
hex trust set autonomous        # Let hex handle routine work
hex taste set error_style=thiserror  # Prefer thiserror over anyhow
hex pulse                       # One-glance system health
hex brief                       # Structured status summary

Three-Path Workplan Dispatch — Local, Remote, and Cloud

The workplan executor classifies every task and routes it through the optimal dispatch path:

hex plan execute workplan.json
  |
  +- Path C (T1/T2/T2.5) --- headless inference --> Ollama (local or remote)
  |   No agent process spawned. Direct inference + GBNF grammar + compile gate.
  |   Fastest path: typo fix in 2.3s, function generation in 10s.
  |
  +- Path A (T3 fallback) --- spawn hex-agent ----> local process with full tooling
  |   For multi-file features that need filesystem access, git, and tool use.
  |
  +- Path B (Claude Code) --- inference queue -----> Claude session dispatches
      When running inside Claude Code, tasks queue for the outer session.

Path C is the breakthrough. It eliminates the agent spawning overhead for 70% of workplan tasks. Instead of forking a process, loading tools, and waiting for a shell — the executor sends the prompt directly to Ollama with a GBNF grammar constraint and gets compilable code back in seconds. The inference router picks the best available server automatically, whether it's localhost or a GPU box on your LAN.