hexum_loop_start
Last updated 3 September 2026
When to call
When starting a gated build loop for a goal.
Arguments
Required goal. Optional max_turns (default 12, max 24), test_command, files.
What comes back
loop_id and the next instruction (shrink, edit, check, test, tick).
What the agent does next
Follow the instruction. Tick with compile and test output.
All five tools: Hexum MCP tools. Contract file: architecture.hexum.
FAQ
What is hexum_loop_start?
Returns loop_id and the next instruction. Then hexum_loop_tick with compile_output and test_output. hexum_loop_status reads; hexum_loop_abort stops. Loops live 15 minutes in memory. ZDR: no payload tables.
When should the agent call hexum_loop_start?
When starting a gated build loop for a goal.
Does Hexum replace Cursor or Claude Code?
No. Hexum is a hosted MCP server. You keep Cursor, Claude Code, Antigravity, Codex, or Grok. Hexum is not an IDE and not a model.
Is hexum_check an LLM?
No. hexum_check is deterministic. It maps files to layers in architecture.hexum and reports forbidden imports. It does not call a model.
Does Hexum store my repo?
No. Zero Data Retention is on Usage and Teams. MCP payloads are not stored. Hexum does not train on your code. Policy: /zdr.