nfemmanuel/iranti

v0.3.6 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 2mo AI Agents & Assistants

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Summary

AI summary

Operating rules now delivered once per context window, reducing overhead.

Full changelog

What's new in v0.3.6

Context overhead optimizations

Operating rules delivered once per context window (not on every handshake) — subsequent calls return a stub until postCompaction:true resets delivery
sessionRecovery suppressed on non-first handshakes within a window
usageGuidance compact path: drops WARNING/NON-COMPLIANT escalation suffix and note when agent is compliant (turnsWithoutWrite < 2)

Cross-LLM protocol enforcement

Critical call sequence embedded in iranti_attend tool description — visible to all LLMs before every tool call decision
Immediate-write timing rule embedded in iranti_write tool description
Protocol now enforced at the tool layer, not just via Claude Code hooks

Attendant operating rules v0.3.3

20 rules → 9 (~60% token reduction)
ACKNOWLEDGE ritual removed
Verbose ATTEND PROTOCOL removed (now in tool descriptions)

SDK

postCompaction?: boolean passthrough in HandshakeInput

View diff on GitHub

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

Share on X Share on Bluesky

Track nfemmanuel/iranti

Get notified when new releases ship.

About nfemmanuel/iranti

Persistent shared memory for AI coding agents. Stores facts as `entity/key/value` triples with hybrid semantic search, task checkpoints, and conflict resolution — shared across Claude Code, Codex CLI, and GitHub Copilot.

All releases →