trueforge
The open-source agent harness - the runtime layer that turns an LLM into a working agent.
Runssource: GitHubhomepageTypeScriptMITcommit 0fe346a7cb09
TypeScript, MIT licensed. The project labels itself: agent, agentic ai, agents, harness and harness engineering.
trueforge runs. An Argusic agent installed it in 13 minutes and hit 4 errors on a clean machine with no GPU, and the whole session was recorded.
TrueForge server starts in standalone SQLite mode on port 8790, responds 200 on /healthz and /api/v1/docs, and lists agents via GET /api/v1/agents. Core and assistant-ui-runtime packages pass all tests. The trueforge package passes 65 of 68 test suites (576 tests) with 3 suites OOM-limited. Frontend build and UI tests are memory-constrained in this container.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 80 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- Frontend (Vite + Monaco editor) build OOM in ~2GB container2 minutes
- 3 trueforge test suites OOM-killed (trueFoundryNaming, perServerMcpHeaders, schedule)
- TrueForge UI test suite partially OOM-killed with some pre-existing test failures
- Local sandbox unavailable: bubblewrap, socat not installed (no root)
- Install time
- 13 minutes
- Cold machine to finish
- 71 minutes
- Errors hit and fixed
- 4 hit, 0 fixed with no human help
- How the result was proved
- curl -s http://localhost:8790/healthz returned {"status":"ok","version":"0.3.0-rc.0"}; curl -s -o /dev/null -w %{http_code} http://localhost:8790/api/v1/docs returned 200; curl -s http://localhost:8790/api/v1/agents returned {"data":[],"pagination":{"limit":50}}; pnpm test:trueforge-core exited 0 with 49 suites, 469 passed; pnpm test:assistant-ui-runtime exited 0 with 22 files, 198 passed; pnpm test:trueforge had 65 suites passed, 576 tests passed, 3 OOM-killed
- Model tokens used
- 248,594
- Exact commit tested
- 0fe346a7cb09
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
The server is fully working. Let me now kill it and produce the final report exec /bin/sh -lc 'pkill -f "tsx.*src/main.ts" 2>/dev/null; echo "killed"' in /work/repo exited 143 in 0ms: Now I have all the information. Let me compile the final result exec /bin/sh -lc 'export PATH="/tmp/node-v22.14.0-linux-x64/bin:$PATH" && cd /work/repo && echo "=== TRU… Now I have all the data. Here's my final report: 2026-09-29T11:26:38.628124Z ERROR codex_core::session: failed to record rollout items: thread 01a0e… tokens used 248,594 Now I have all the data. Here's my final report:
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
What did not, or is not known
- Took 13 minutes to install, slower than the median of the 56 comparable projects Argusic has measured.
- Hit 4 errors during setup, 4 of which were not solved.
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
Of the 57 agent projects Argusic has installed and timed, trueforge was the 39th fastest to reach a running state, and 39 of 57 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/trueforge)Questions
- Does trueforge run?
- Yes. trueforge runs. Argusic installed and launched it on a clean machine in 13 minutes, hitting 4 errors on the way, and recorded the session.
- How did Argusic test trueforge?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 0fe346a7cb09. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test trueforge?
- The run that produced this verdict cost $0.12: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does trueforge take to install?
- 13 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does trueforge need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing trueforge?
- 4 things broke in the recorded run. Each one, and the time it cost, is listed on this page.
- How does trueforge compare with the alternatives?
- Of the 57 agent projects Argusic has installed and timed, trueforge was the 39th fastest to reach a running state, and 39 of 57 reached one at all.
- Where is the evidence for trueforge?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.