New Checked the same way as MCP servers
Agent Skills, and what they let an agent do
Read straight from each publisher's repository. For every skill: what it allows the agent to do, the scripts it ships, and what fixed checks matched. Nothing is run.
Skills from OpenAI3
Clear filters- migrate-to-codexOpenAI · 15 scriptsMigrate supported instruction files, skills, agents, and MCP config into Codex project and global files.OfficialDestructive commandsRuns commands1 high finding
- skill-installerOpenAI · 3 scriptsInstall Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).OfficialDestructive commandsReaches the network1 high finding
- vercel-deployOpenAI · 1 scriptDeploy applications and websites to Vercel. Use when the user requests deployment actions like "deploy my app", "deploy and give me the link", "push this live", or "create a preview deployment".OfficialDestructive commandsReaches the network2 high findings
Labels come from the skill's own files, read on the date shown: the tools its frontmatter allows and the scripts it ships. They are not a safety rating, and no finding is not a safety guarantee. How skills are checked.