ilo 0.12.0

ilo - the token-minimal programming language AI agents write
Documentation
/target
npm/ilo.wasm
.env
__pycache__/
/ilo
*.png
.playwright-mcp/
.worktrees
.claude/worktrees/
# research/ is untracked by default (scratch space + private explorations).
# Track the closed-loop benchmark explicitly - it is a publishable artefact.
research/*
!research/closed-loop-bench/
!research/BENCHMARK-METHODOLOGY.md
# PNGs in the benchmark charts/ are dataset artefacts and must be tracked.
!research/closed-loop-bench/charts/*.png
# Per-cell raw results and tool caches are reproducible from the driver -
# don't bloat the repo with them.
research/closed-loop-bench/results/raw/
research/closed-loop-bench/.zero/
research/closed-loop-bench/**/__pycache__/
site/
.DS_Store
.astro/
node_modules/
rust_out

# Research / scratch docs (subagents leave reports here; not tracked)
docs/

# E2E test artefacts (subagent-produced logs; not tracked)
tests/e2e/

# CLAUDE.md is per-developer Claude Code context; AGENTS.md is the cross-tool tracked equivalent
CLAUDE.md

# Internal strategy and exploration (kept local, not public)
zero-gap-specs/
research/
.claude/commands/
.env.example
.zero/