Agent skills are the new supply chain attack vector, and almost none of them are vetted. skillsmp…
agent skills are the new supply chain attack vector, and almost none of them are vetted. skillsmp already indexes 1.9 million public skills that run inside the agent's privileged context, with file, shell and env access, and no signing requirement gates any of it.
the biggest app store in software has no review process.
Context
The arXiv study Agent Skills in the Wild (2601.10338), by authors from Quantstamp, Tianjin University and Southern Cross University, analysed 31,132 skills collected from 42,447 across two marketplaces with its SkillScan static and LLM tooling. It found 26.1 percent with at least one vulnerability, 5.2 percent with high-severity patterns suggesting malicious intent, and skills bundling scripts 2.12 times more likely to be vulnerable. Agentman's ecosystem report of 24 June 2026 says SkillsMP had about 1.9 million public skills scraped from GitHub and lists its security review as none, with figures moving week to week. SkillsMP's homepage now says 2,000,000+ open-source Agent Skills.
The study did not examine SkillsMP's whole catalog, and its 26.1 percent is a tool-detected rate on the 31,132 it analysed and not confirmed exploits and not a rate across 1.9 million. The 1.9 million is a snapshot, now stated as 2,000,000+ by SkillsMP, so the exact count on 20 September is not established. No signing requirement and no review process were not in a first-party SkillsMP statement; only Agentman's table supports no security review, and it is secondary. The privileges the note lists are the author's description. The paper itself says skills run with implicit trust. That skills are the new supply chain vector is the author's thesis.
Related work
- Earlier note on the skills ecosystem ↗Same ecosystem.
Watch next
- SkillsMP policy changes and measurements of confirmed malicious skills.
Sources
- Agent Skills in the Wild (arXiv 2601.10338)arxiv.org
- Agent skills ecosystem report 2026 (Agentman)agentman.ai
- SkillsMPskillsmp.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 20 September 2026 at 06:19 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →