This release adds 1 notable feature for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+2 more
Summary
AI summaryAdded optional TOON encoding for JSON tool results reducing token usage by ~60% on large uniform data.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
Adds optional `/toon` endpoint to re-encode JSON results into TOON format. Adds optional `/toon` endpoint to re-encode JSON results into TOON format. Source: llm_adapter@2026-06-07 Confidence: high |
— |
| Performance | Low |
TOON encoding reduces token count by ~60% on large uniform data. TOON encoding reduces token count by ~60% on large uniform data. Source: granite4.1:30b@2026-06-07-audit Confidence: low |
— |
Full changelog
Added
/toon(opt-in) — re-encode JSON tool results into TOON (Token-Oriented Object Notation): an array's field names are written once like a CSV header instead of repeated on every row, and scalar arrays are inlined. On large uniform data that's ~60% fewer tokens than JSON (a 50-row array: 788 → 281 tokens). It only swaps in TOON when the result is genuinely shorter (never makes output larger), and non-JSON results pass through untouched. Pairs well with MCP servers that return big JSON payloads.
Tests
test_toon.py— encoder (tabular/scalar arrays, nesting, JSON fallback, quoting, numeric-string preservation) andmaybe_toon(savings, never-longer, pass-through). 443 total; ruff + CI green.
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Kodiqa Agent
All releases →Related context
Related tools
Beta — feedback welcome: [email protected]