Argusic daily report, 13 October 2026

On 13 October 2026 (UTC) an AI agent tried to install and run 57 open-source projects on clean machines, with nobody helping: 63 attempts in all, 4 of them cut short by our own tooling and not counted.

Verdicts as they stood at the end of that day. A project tested again since may have moved; its own page always shows the current verdict. Written 2026-10-14.

The day in numbers

Projects: 57
Attempts: 63
Runs: 28
Runs with mocks: 21
Could not verify: 3
Not yet verified: 5

Five from the day

Chosen by a fixed rule, not by hand: the three most starred projects of the day, then the most starred one that runs, then the most starred one that could not be verified.

Fabric44,248 stars, runs
Unlimited-OCR27,295 stars, runs with mocks
gridcraft4,400 stars, not yet verified
fae852 stars, runs
dbos-transact-golang853 stars, could not verify

All 57 projects, most starred first

subject	verdict	topic	stars	attempts	evidence
Fabric	Runs	ai	44,248	1	log and recording
Unlimited-OCR	Runs with mocks	Python	27,295	1	log and recording
gridcraft	Not yet verified	Rust	4,400	1	log and recording
cadcraft	Not yet verified	Rust	4,132	1	log and recording
dbos-transact-golang	Could not verify	automation-framework	853	2	log and recording
concretecms	Runs with mocks	cms	852	1	log and recording
fae	Runs	cms	852	1	log and recording
agentfiles	Runs	agent-skills	851	1	log and recording
deco	Runs with mocks	cms	850	1	log and recording
Scavenger	Runs with mocks	bot	847	1	log and recording
NazaraEngine	Could not verify	3d-engine	843	2	log and recording
hue	Runs with mocks	ai-tools	842	1	log and recording
whiteboard	Runs	collaborative	842	1	log and recording
golf	Runs with mocks	agent-runtime	840	1	log and recording
yokai	Runs	backend	840	1	log and recording
motor-admin-rails	Could not verify	admin	839	2	log and recording
temps	Runs	ai-agents	839	1	log and recording
ue4-docker	Runs with mocks	continuous-integration	839	1	log and recording
versus-incident	Runs	ai	839	1	log and recording
litiengine	Runs	2d	838	1	log and recording
search_cop	Runs	activerecord	838	1	log and recording
memorix	Runs	agent-memory	836	1	log and recording
spidr	Runs with mocks	crawler	836	1	log and recording
gemini-skill	Runs with mocks	automation	834	1	log and recording
sleepless-agent	Runs with mocks	ai-agent	834	1	log and recording
supabase-mcp-server	Runs	cursor	831	2	log and recording
mcp-client-for-ollama	Runs with mocks	agentic-ai	830	1	log and recording
lazywarden	Runs with mocks	backup	828	1	log and recording
lx-netease-music-mobile	Runs with mocks	music-player	826	1	log and recording
scira-mcp-chat	Runs	anthropic	825	1	log and recording
Eldiron	Runs	classic	824	1	log and recording
gateway	Runs with mocks	api	824	1	log and recording
automem	Runs with mocks	ai	822	1	log and recording
jenkins-library	Runs	ci-cd	822	1	log and recording
velite	Runs	content	822	1	log and recording
semantix	Runs	agent-framework	817	1	log and recording
blades	Runs with mocks	agent	816	1	log and recording
blinkpy	Runs with mocks	api	813	1	log and recording
dsh-browser	Runs	browser-automation	813	1	log and recording
keifu	Runs	cli	811	1	log and recording
tokentap	Runs	ai-tools	811	1	log and recording
vector-db-from-scratch	Runs	apache-arrow	807	1	log and recording
seonaut	Runs	crawler	806	1	log and recording
lobe-tts	Runs with mocks	auzre	805	1	log and recording
os	Runs with mocks	agentic-ai	805	1	log and recording
jvppeteer	Not yet verified	chrome	804	1	log and recording
luya	Runs	admin	802	1	log and recording
second-brain-cloudflare	Runs	ai-memory	802	1	log and recording
jadx-mcp-server	Runs with mocks	ai	801	1	log and recording
agent-console	Not yet verified	ai-agents	800	3	log and recording
aralez	Runs	load-balancer	800	1	log and recording
claude-code-zh-cn	Runs with mocks	chinese	800	1	log and recording
rtm	Runs	c-plus-plus	800	1	log and recording
nestjs-otel	Runs	hacktoberfest	799	1	log and recording
sharp	Runs	admin-dashboard	799	1	log and recording
core	Runs	brokers	798	1	log and recording
patchright-nodejs	Not yet verified	automation	796	1	log and recording

How a verdict is reached: the methodology. Every project tested so far: the subjects. GPU cloud and VPS prices: servers. Laptops and desktops for programming and AI: hardware. Earlier days: the daily archive.