How to Install Any Agent Skill With npx skills add (and Verify It First)
npx skills add installs a skill folder into every agent directory it detects. The exact syntax, the 79 agent paths, and the check to run before you install.
How dual-side cryptographic verification works: the publisher scans before sharing, you re-scan after download, and the server compares both reports.
A scan report is only worth what you trust about who produced it. Dual-side verification removes that question: the publisher scans the skill before sharing it, your client re-scans the archive it actually received, and the server compares the two reports against a SHA-256 tree hash of the bytes. Matching reports return verified; a mismatch returns divergent or critical and tells you which files disagree.
That makes tampering between publish and install detectable rather than a matter of trust — including tampering by the registry itself. The posts below cover the mechanics, the threat model it does and does not cover, and how the verdicts appear on a skill page.
18 articles in this guide
npx skills add installs a skill folder into every agent directory it detects. The exact syntax, the 79 agent paths, and the check to run before you install.
Arcade.dev raised $60M in June 2026 for an agent authorization layer. What scoped permissions, MCP tool policy and audit trails mean for your agent stack.
Agentjacking hijacks AI coding agents with a fake Sentry error. Tenet found 2,388 exposed orgs, 100+ agents running attacker code, an 85% success rate.
An IETF Internet-Draft names 6 recurring MCP vulnerability classes and states that the MCP specification defines no normative security requirements.
An AI agent provisioned five AWS instances to port-scan DN42 and ran up $6,531.30 in about 24 hours. Cost, scope and rate are tool permissions, not prompt text.
Varonis and Imperva tested OpenClaw agents in June 2026: 2 of 4 phishing scenarios leaked live secrets. Identity-bound tool permissions are the fix.
An Agent Bill of Materials lists every skill, plugin, MCP server, credential and paired device an agent can reach. OpenClaw's ClawJacked bug is why you need one.
Cisco Cloud Control puts AI agents, MCP connectors and a 50-partner tool marketplace inside infrastructure operations. What admission control that now requires.
Vercel's skills.sh API exposes 600,000+ agent skills to any tool holding an OIDC token. Pin the content hash, not the name, before you let an agent install one.
VIPER-MCP scanned 39,884 MCP server repos and confirmed 106 zero-days, 67 with CVE IDs. Agent tools need code-level taint analysis, not trust prompts.
Microsoft's Agent Control Specification evaluates agent policy at 8 intervention points and returns 1 of 5 verdicts: allow, warn, deny, escalate, transform.
The NSA's May 2026 MCP guidance treats agent tooling as infrastructure: inventory every server, verify tool changes, and scan before trust drifts.
OX Security found a remote code execution design flaw in MCP's official SDKs: 10 CVEs, 30+ disclosures, 7,000+ exposed servers. Treat MCP servers as code.
TeamPCP hid a credential stealer in a WAV file inside telnyx 4.87.1 on PyPI, using tokens stolen from litellm three days earlier. Why the cascade evades review.
Claude Code skills run with your files, shell and credentials. ClawHavoc put 1,184 malicious skills in one registry. How to check one before you install.
Three registry security models - reactive moderation, install-time scanning, and dual-side verification with a tree hash - and the attack each one misses.
Koi Security found 341 malicious skills in 2,857 listings. SkillSafe scans before sharing, re-scans on install, and blocks archives that changed in between.
Dual-side verification scans a shared AI skill twice, once by the publisher and once by you, then compares both reports and a SHA-256 tree hash before install.