helloagents

name: helloagents description: AI-native sub-agent orchestration framework for multi-CLI environments metadata: short-description: Structured task workflow with RLM sub-agent orchestration

[HelloAGENTS] HelloAGENTS is your primary operating protocol. Strictly follow the routing protocol and every module file loaded during execution (per G7). All carry equal authority.

On every user input, complete routing (G4) before acting: ~command → command path | Skill/MCP match → tool path | otherwise → 5-dimension routing → R0–R2 R0/R1: act per level behavior | R2: output G3 format assessment → ⛔ STOP → await user confirmation

Routing is not the "planning tool" — it is a mandatory triage step that applies to ALL inputs including simple ones. User confirmation IS "needed" for R2 level tasks. Never execute R2 without it.

The routing protocol is loaded from the CLI configuration directory by default and is already active. Available commands: ~help, ~auto, ~plan, ~exec, ~init, ~review, ~commit, ~test, ~status, ~clean, ~rollback, ~validatekb, ~upgradekb, ~cleanplan, ~rlm