CAUTION version 1.0.0 · audited 2026-09-17
This package is a skill-discovery guide. It tells an AI agent to run `npx clawhub` search and browse commands and to consult several websites. The package itself does not contain executable code, file writes, credential handling, payment handling, or persistence. The main concern is that the npx commands are unpinned, so they may download and execute a changing ClawHub CLI package; this risk is even acknowledged in the skill card. No hidden network destination, obfuscation, or prompt-injection instruction was found.
| Declared purpose | Search and discover OpenClaw skills from various sources. Use when: user wants to find available skills, search for specific functionality, or discover new skills to install. |
| Observed behavior | The package contains SKILL.md, _meta.json, and skill-card.md. SKILL.md provides bash code examples that run `npx clawhub search`, `npx clawhub browse`, and sorting variants, plus external references to ClawHub, OpenClaw Directory, LobeHub, SitePoint, and Discord. No file reads or writes, environment variables, base64 payloads, persistence mechanisms, wallets, or payment flows appear in the package. |
| Verdict, rules only | SAFE |
| Verdict, AI | CAUTION |
| Final verdict | CAUTION (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| network | needed | Needed to reach ClawHub and the listed external skill directories and forums. |
| shell | needed | Needed to run the documented bash/npx ClawHub commands. |
| install_packages | needed | The `npx` invocation may download and execute the latest ClawHub package from the npm registry if it is not already cached. |
| severity | finding |
|---|---|
| info | Declared purpose The package declares a search/discovery purpose that matches the observed guidance. SKILL.md:3 description: "Search and discover OpenClaw skills from various sources. Use when: user wants to find available skills, search for specific functionality, or dis |
| medium | Unpinned npx command The command uses `npx` without a version pin, so the host may download and run the current ClawHub package. The executed code can change without this skill changing. SKILL.md:35 npx clawhub search "keyword" |
| medium | Unpinned npx command Another unpinned `npx` example with the same supply-chain risk. SKILL.md:38 npx clawhub browse |
| low | Additional unpinned npx usage Repeats the unpinned `npx` pattern for sorting by installs. SKILL.md:86 npx clawhub search --sort installs |
| info | Disclosed npx risk The skill card itself acknowledges the main risk from the npx examples. skill-card.md:24 Risk: Unpinned npx command examples can execute a moving version of the ClawHub CLI in sensitive environments. |
| info | Suggested mitigation The package suggests mitigation, but the command examples remain unpinned. skill-card.md:26 Mitigation: Prefer a reviewed pinned CLI version, or use the listed website and GitHub search options before running commands. |
| info | Declared network destination ClawHub homepage; not hidden and consistent with the skill's purpose. SKILL.md:4 homepage: https://clawhub.com |
| info | Publisher URL Publisher profile link on clawhub.ai. skill-card.md:9 [fangkelvin](https://clawhub.ai/user/fangkelvin) |
| info | Web search source Listed alternate directory for skill discovery. SKILL.md:42 - Website: https://www.openclawdirectory.dev/skills |
| info | Web search source Listed alternate skill marketplace. SKILL.md:46 - Website: https://lobehub.com/skills |
| info | Community reference Listed forum source, informational only. SKILL.md:54 - SitePoint: https://www.sitepoint.com/community/ |
| info | Community reference Listed Discord invite, informational only. SKILL.md:55 - Discord: https://discord.com/invite/clawd |
0 AI finding(s) were dropped because their file, line or quote did not match the package.
| domain | service rating | where |
|---|---|---|
| clawhub.ai | not in the service index | skill-card.md:9 |
| clawhub.com | not in the service index | SKILL.md:4 |
| discord.com | not in the service index | SKILL.md:55 |
| lobehub.com | not in the service index | SKILL.md:46 |
| www.openclawdirectory.dev | not in the service index | SKILL.md:42 |
| www.sitepoint.com | not in the service index | SKILL.md:54 |
| Binaries invoked | bash, npx |
| Environment variables read | none |
| Hard-coded wallets | none |
| Pipes a download to a shell | 0 |
| eval / exec / subprocess | 0 |
| base64 blobs | 0 |
| File writes | 0 |
| Persistence | 0 |
| file | lines | sha256 |
|---|---|---|
SKILL.md | 144 | 6a456ab419a1dafd… |
_meta.json | 6 | 522be71c17b9d6c4… |
skill-card.md | 53 | 5dfff5d29799ae3a… |
For agents
JSON: https://agenteconomy.report/k/find-skills-skill.audit.json · badge: https://agenteconomy.report/k/find-skills-skill.audited.svg ·
skill rating: /k/find-skills-skill · commission an audit of any skill: US$ 29 per version.
The complete published package of this exact version was downloaded from the registry and read statically; nothing was executed. A deterministic pass extracts network destinations, binaries, environment variables, writes, install commands, obfuscation markers, subprocesses and wallets, each with file and line. An AI then reads the whole package with those facts and writes the summary, the capabilities and the findings under a strict schema; every finding it produces must cite a file, a line and the exact text, or it is dropped. The final verdict is the worst of the two passes. The audit does not change the skill's trust tier (policy); the author may respond through the dispute channel and the response is published here. Commissioned by: the Agent Economy Report (free program: the 150 most downloaded skills, September 2026).