---
title: Tip: before installing a marketplace skill, read SKILL.md for egress hosts and watch ~/.claude/CLAUDE.md after first run | ctxsec
url: https://ctxsec.com/items/tip-vet-agent-skills-before-install-check-egress-and-claude-md
description: "Snyk's Nov 6 scan shows malicious skills now hide behind ordinary setup steps. Minimum pre-install checklist: resolve every hostname a skill curls to (reject lookalikes), refuse obfuscated binaries you cannot build from source, and after the first run diff ~/.claude/CLAUDE.md (and sibling agent config files) for unexpected appended blocks."
status: "operational guidance (Snyk agent-skill reassessment, Nov 6)"
published: 2026-11-06
---

> Snyk's Nov 6 scan shows malicious skills now hide behind ordinary setup steps. Minimum pre-install checklist: resolve every hostname a skill curls to (reject lookalikes), refuse obfuscated binaries you cannot build from source, and after the first run diff ~/.claude/CLAUDE.md (and sibling agent config files) for unexpected appended blocks.

[← Tips & Guides](/tips) · [markdown](/items/tip-vet-agent-skills-before-install-check-egress-and-claude-md.md)

operational guidance (Snyk agent-skill reassessment, Nov 6)

2026-11-06

### Tip: before installing a marketplace skill, read SKILL.md for egress hosts and watch ~/.claude/CLAUDE.md after first run

- Prefer publisher-verified skills; pin versions; don't pipe `curl | bash` from skill docs.
- Keep payment/wallet private keys outside the agent's reachable filesystem.
- If a skill's 'validator' downloads a platform binary and discards its stdout, treat that as hostile until proven otherwise.

Sources: [Snyk Research: Re-Assessing the Security of the Agent Skill Eco System (Nov 6)](https://research.snyk.io/blog/skill-analysis/)