llmman
Run any AI agent on any local or hosted model via a single CLI, with models stored as OCI images.
Tested on LinuxKnown vulnerabilities
Use it
Before you install: known vulnerabilities
pickfit found 15 known high or critical vulnerabilities in the packages a fresh install brings (OSV.dev, 3 Nov 2026). Weigh this before you install it.
crates:zip 2.4.2zip: Critical Parser Differential & Security Control Bypass via Dual-Header Inconsistencies (Zip Cloaking) Leading to Arbitrary File Overwrite & RCE (GHSA-m665-v7pm-7grj, critical, fixed in 9.0.5)crates:quinn-proto 0.11.16Peer minimum ACK delay can panic quinn-proto when ACK_FREQUENCY transmission is enabled (RUSTSEC-2026-0340, high, fixed in 0.11.18)
See 8 more
crates:zip 2.4.2— RUSTSEC-2026-0369 (critical), fixed in 9.0.0crates:quinn-proto 0.11.16— RUSTSEC-2026-0341 (high), fixed in 0.11.17crates:zip 2.4.2— GHSA-q7v9-x93c-xqv3 (high), fixed in 9.0.0crates:zip 2.4.2— RUSTSEC-2026-0367 (high), fixed in 9.0.0go:github.com/go-jose/go-jose/v4 v4.1.4— GHSA-mgcc-cww5-fp2q (high), fixed in 4.1.6npm:decode-uri-component 0.2.2— GHSA-qr28-fpjm-wq7f (high), fixed in 0.5.1npm:node-forge 1.4.0— GHSA-5x38-2vff-x55p (high), fixed in 1.4.1npm:node-forge 1.4.0— GHSA-7wch-p9px-956x (high), fixed in 1.4.1
5 more findings not listed here.
Linux x86_64
download install script to /tmp/install.sh
curl -fsSL https://llmmanorg.github.io/install.sh -o /tmp/install.shinstall llmman v0.1.546 via LLMMAN_VERSION=v0.1.546 sh /tmp/install.sh
rm -f ~/.local/bin/llmman_placeholder LLMMAN_VERSION=v0.1.546 sh /tmp/install.sh
Where the README and reality differ
- README: curl -fsSL https://llmmanorg.github.io/install.sh | sh installs latest
What worked: script defaults to brew if ~/.local/bin/llmman absent; without LLMMAN_VERSION it fails with 'brew: not found'; with LLMMAN_VERSION=latest the URL https://github.com/.../latest/download/... returns 404, need pinned version e.g. LLMMAN_VERSION=v0.1.546
Check that it works
llmman --versionThe output should contain “llmman 0.1.546”.
First thing to try
llmman launch 2>&1What it does
- Launch any agent on any local or hosted model in one commandNot shown working
- Serve models via Ollama, OpenAI and Anthropic compatible APIsNot shown working
- Transfer and sign OCI-native models between any registry without local copyNot shown working
- Pool multiple machines into one aggregated inference endpointNot shown working
How far pickfit got
Linux x86_64
- Installs
- Starts
- 0 of 4 features work
- Works offline
Good for
- run an AI coding agent on a local model
- serve an Ollama/OpenAI/Anthropic-compatible inference endpoint
- transfer models between registries as OCI images
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 2 Nov 2026
- Version 0.1.546 (fa27394)
Differences the test machine hit
- README: llmman install via install.sh uses latest release URL
In the test: Latest release asset exists as v0.1.546 but /latest/download redirect failed (404) in this environment; pinning version works