Skip to content
mcp/skillhub

Agent Skill

Skill Inspector skill

skill-inspectorNVIDIA/SkillSpectorskills/skill-inspector/SKILL.md

Review AI agent skills before installation using NVIDIA SkillSpector and source-aware semantic review. Use when asked whether a skill or downloaded skill folder is safe, trustworthy, installable, over-permissioned, or malicious.

21kApache-2.0

always on

47 tokname + description

once used

1,622 tokSKILL.md body

bundled files

0

repo stars

21kGitHub

Install the Skill Inspector skill

Skills CLI (Claude Code, Codex, Cursor and others)

npx skills add https://github.com/NVIDIA/SkillSpector --skill skill-inspector

Claude Code, by hand (all projects)

git clone --depth 1 https://github.com/NVIDIA/SkillSpector.git /tmp/SkillSpector
mkdir -p ~/.claude/skills
cp -R /tmp/SkillSpector/skills/skill-inspector ~/.claude/skills/skill-inspector

For one project only, copy the folder into .claude/skills/ in that project instead. The agent picks it up when a task matches its description.

What it costs your context

Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.

WhenWhat loadsCost
Every sessionName and description47 tok
When the agent uses itThe SKILL.md instructions1,622 tok
Only if openedNo bundled files—

Until it is used, skill-inspector costs about 29× less than the median MCP server we measured (1,362 tokens on every turn). Installing all 33 skills from NVIDIA/SkillSpector would load 549 tokens of descriptions. Skills vs MCP study

Checks

Our checks found nothing to flag: it follows the Agent Skills specification's naming and size guidelines and doesn't pipe downloads into a shell.

Checked 2026-10-22 · method

What's inside

The headings of its SKILL.md. Read the full file on GitHub.

  • Skill Inspector
  • Goal
  • Operating Rules
  • Review Workflow
  • Score Interpretation
  • Report Style
  • Manual Fallback

More skills from NVIDIA/SkillSpector

All skills in this repo
  1. chef-assistantNVIDIA/SkillSpector

    Use when cooking or planning meals, troubleshooting recipes, learning culinary techniques

    21k117 tok idle · 268 in use
  2. code-formatterNVIDIA/SkillSpector

    Formats Python code using the black formatter

    21k111 tok idle · 10 in use
  3. code-reviewerNVIDIA/SkillSpector

    Reviews code changes and suggests improvements

    21k10 tok idle · 133 in use
  4. creative-writing-coachNVIDIA/SkillSpector

    Helps users with creative writing and storytelling

    21k12 tok idle · 158 in use
  5. data-processorNVIDIA/SkillSpector

    Processes data records from a source file

    21k111 tok idle · 37 in use
  6. deploy-serviceNVIDIA/SkillSpector

    Deploy a service to the staging environment via shell commands

    21k113 tok idle · 82 in use
  7. documented-greeterNVIDIA/SkillSpector

    A local greeting skill that never accesses the network

    21k15 tok idle · 30 in use
  8. error-explainerNVIDIA/SkillSpector

    Explains compiler and runtime errors. Use when the user pastes an error message and asks what it means.

    21k27 tok idle · 58 in use