coldpass

google-research/harness2

GitHub ↗ · How this check works

Per-setup public labels: Linux / Haiku: Works

Static status badge · SVG · read-only JSON. Embed with ![Coldpass status](https://coldpass.dev/badge/google-research--harness2.svg).

Works

Only older checks are saved; no current product test.

Answers apply to each recorded setup, revision and operation. Earlier observations remain in Run details.

Tested on Linux
PlatformHaikuLuna
Linuxyes

Linux / Haiku 5.5: yes · Oct 18, 2026 · 1cfe254e8d25 · claude-code / 2.1.300 (Claude Code)
Recorded operation: /workspace/.venv/bin/harness2-demo --output runs/demo --k 2

Needs keys or browser
None seen
Setup time
5 s to install, 4 s to first use
Issue
none found in recorded check

Safety: no observations were recorded for the selected revision. Functional results do not establish a safety result.

Run details

Haiku 5.5

Native agent · Checked operation passed

Checked operation /workspace/.venv/bin/harness2-demo --output runs/demo --k 2

Agent-reported scope: Local deterministic demo: harness2-demo --output runs/demo --k 2 (both recursion modes with real Python executions; no model credentials). Completed 2 tasks, 0 failed, wrote runs/demo/demo.json.

Run details
run
google-research--harness2-claude-haiku-5-5-1792335890898464886
date
Oct 18, 2026
time to working
35.9s
product hosting / execution
Local command in the guest
recorded operation scope
product operation
model
claude-haiku-5-5
agent / version
claude-code / 2.1.300 (Claude Code)
platform
linux
reported cost
$0.004521
Recorded source
https://github.com/google-research/harness2
Tool transport
mcp-assigned-vm
Test version
readme-d82246e0bbc037649a4db516d4aa456c0500644302d6c13970f8d8766d8bd681
revision
1cfe254e8d25b15f37048c3730134686abe60335
Whose fault: not assigned · resolved

What the agent did

Local deterministic demo: harness2-demo --output runs/demo --k 2 (both recursion modes with real Python executions; no model credentials). Completed 2 tasks, 0 failed, wrote runs/demo/demo.json.

curl -LsSf https://astral.sh/uv/install.sh | sh (installed uv 0.13.1; guest lacked uv and pip)
uv python install 3.12 (cpython 3.12.15)
uv venv --python 3.12 .venv
source .venv/bin/activate && uv pip install -e '.[render]' (installed harness2 0.1.0 and deps)
harness2-demo --output runs/demo --k 2 (rc=0: 2/2 ledger tasks ok, 0 failed)
removed generated runs/ so the checker gets a fresh output path

run JSON · checker evidence · agent report + check · transcript excerpt · error excerpt