Skip to content

Hollow

v3.14.1 Feature

This release adds 2 notable features for engineering teams evaluating rollout.

✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Summary

AI summary

qwen3.5:9b reasoning model fixed to return proper response fields instead of empty params.

Full changelog

What's new

Agents now have a live thought stream you can watch in real time.

Real-time thoughts log (/agentOS/logs/thoughts.log)

Every agent broadcasts its inner monologue as it works:

  • GOAL: — the objective it's pursuing
  • PLAN: — the full capability chain Ollama chose
  • step N: — each step with its generated params
  • RUN/OK/FAIL: — execution result previews as they happen

Watch it live:
```
wsl -e bash -c "tail -f /agentOS/logs/thoughts.log"
```

qwen3.5:9b thinking mode fix

qwen3.5:9b is a reasoning model — it was putting all output in a thinking field and leaving response empty, causing every capability to run with empty {} params. Fixed by adding "think": false to all Ollama API calls, which disables chain-of-thought mode and returns clean structured JSON.

Agents are now generating real parameters and making actual progress on their goals.

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.

Sign up free

About Hollow

All releases →

Related context

Earlier breaking changes

  • v5.7.32 Web dashboard removed; operator panel is canonical UI

Beta — feedback welcome: [email protected]