Browse / Agent tools
loadn-ai/loadn
Self-contained coding-agent engine (Claude Code stream-json dialect, single httpx dep) + multi-agent platform with a 7-layer defense-in-depth security stack. Companion artifact for 3 preprints. Early access; official release before 2026-10-10.
- security
Works.Tested on 2026-10-28. From the README alone, on Linux with Haiku and Linux with Luna.
What we checked
Does it run from the README?Yes. Running in 2:02 on a clean Linux box.The run
Tested Oct 28, 2026, rev 544915a.
Does it do what the README says?1 of 1 claims held up.Evidence
“Headless one-shot prompt prints the agent reply to stdout”
Expected reply text printed. Got LOADN_E2E_OK_7731.
Which setups did we try?Works with Haiku, Luna.Setups
Linux / Haiku: works, Oct 28, rev 544915a. Linux / Luna: works, Oct 28, rev 544915a.
Needs keys, a login or a browser?No.Details
None seen
Is it safe to run?No unexpected network or file activity found in these checksWhat we saw
No problems traced to this repository code in these checks; a short run cannot prove safety. 34 listed items (named-host connections, static matches, dependency advisories) are below and do not change this verdict. 10 static matches retained as context.
OSV reports a known vulnerability in this dependency version. 11 distinct locations. 17 findings.
Reviewed as expected for a test run: attempted writes that failed, so nothing was written.
Full event, path and process lists for each observed setup:
A quiet run is not a safety guarantee.
Is it maintained?Maintained · 553 commits in 90 days · 0 merged PRs in 90 days · 0 open issues · 10 open PRsActivity
actively maintained: last commit 0 days ago; 553 commits and 0 merged PRs in 90 days; 0 open issues, 10 open PRs; latest release 2026-10-28
Can I show this on my README?Copy the badge.Badge
The run
Oct 28 · rev 544915aIndependent test, not endorsed by the project.