Skill Hunter

Find, evaluate, and install ClawHub skills. Semantic search across 10,000+ skills, security vetting before install, side-by-side comparison. The skill that m...

Security Scan

Skill Hunter’s claims match its instructions and requirements: it is an instruction-only discovery/vetting helper that uses ClawHub public APIs and requests no credentials or installs.

Purpose & Capability

Name/description (search, vet, install helper) align with the files and instructions: all example commands call ClawHub public endpoints and the manifest declares network permission. No unexpected credentials, binaries, or config paths are requested.

Instruction Scope

SKILL.md and instructions.md confine actions to reading ClawHub endpoints, inspecting remote SKILL.md files, and advising human-driven install steps (clawhub install). There are no instructions to read local secrets, run arbitrary local scripts, or exfiltrate data beyond search queries to clawhub.ai.

Install Mechanism

This is an instruction-only skill with no install spec and no code files to execute; lowest-risk installation model (nothing is written to disk by the skill itself).

Credentials

No environment variables, credentials, or config paths are required. The guidance describes how to evaluate other skills that might request credentials, but Skill Hunter itself does not request them.

Persistence & Privilege

always is false and the skill does not request elevated or persistent system privileges. It only documents human-driven use of the ClawHub CLI and API; it does not modify other skills or global agent settings.

Assessment

This skill appears coherent and low-risk: it only queries ClawHub public APIs and provides vetting guidance. Before using it, confirm you trust the clawhub.ai domain and your network policy for allowing outbound requests. When following its recommendations to inspect or install other skills, always: (1) read the target skill's SKILL.md and file list, (2) refuse skills that ask you to run 'curl | bash' or install packages from unknown hosts, and (3) avoid granting credentials unless you understand exactly why they are needed. If you want extra assurance, verify the author identity and inspect downloaded skill files locally before running them.

10,000+ skills on ClawHub. No good way to find the right one, vet it, or know if it's safe to install. Skill Hunter fixes that.

Three modes

Hunt — You know what you need. Describe it in plain English. Skill Hunter searches ClawHub's semantic index and returns ranked matches.

curl -s "https://clawhub.ai/api/v1/search?q=code+review+security" | jq '.[] | {slug, displayName, summary, score}'

Scout — You're exploring. Browse trending, newest, or most-downloaded skills across the platform.

curl -s "https://clawhub.ai/api/v1/skills?sort=trending&limit=10" | jq '.items[] | {slug, displayName, summary}'

Vet — You found a skill. Before installing, read its SKILL.md remotely, check its security profile, and understand what it will do on your machine.

curl -s "https://clawhub.ai/api/v1/skills/kenoodl-synthesis/kenoodl-synthesis/file?path=SKILL.md"

Full workflow, API reference, and security evaluation framework in instructions.md.

Security profile

No credentials required. No env vars. No external packages. All API calls go to clawhub.ai — the platform's own public endpoints. Nothing leaves your environment except search queries to ClawHub.