Hollow

v3.14.0 Feature

This release adds 4 notable features for engineering teams evaluating rollout.

Published 2mo AI Agents & Assistants

✓ No known CVEs patched

✓ No known CVEs patched in this version

Summary

AI summary

The autonomy loop now passes step results as context and marks goals completed at progress >= 1.0.

Full changelog

First release where the autonomy loop actually works end-to-end.

autonomy_loop.py: Accepts reasoning_layer (was silently dropped before — the daemon was running blind). Each step passes its result as context to the next. Goal marked completed at progress >= 1.0.
reasoning_layer.py: Improved Ollama prompt with capability descriptions, param format examples, and selection rules.
execution_engine.py: Filters params to only those the function accepts — Ollama-hallucinated extras no longer crash execution. Capabilities returning ok=False now correctly marked failed.
live_capabilities.py: Explicit JSON param schemas. memory_set accepts non-string values. memory_get returns ok=False on error.

Agent receives goal → Ollama selects capability → executes with real params → progress tracked → goal marked completed at 1.0

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

Track Hollow

Get notified when new releases ship.

About Hollow