Agent Reach
WarnAudited by ClawScan on May 10, 2026.
Overview
Agent Reach is a broad web and social-media skill, but it asks the agent to use account cookies, bot-bypass tooling, undeclared external helpers, and posting commands that need careful review.
Review this skill carefully before installing. Use dedicated API keys and isolated browser profiles, avoid giving it your main browser cookies, do not run unreviewed local helpers, avoid anti-bot bypass workflows unless authorized, and require explicit confirmation before any public posting or account action.
Findings (6)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Your logged-in social or browser sessions could be used by the agent or external tools, exposing account data or enabling account actions.
This asks the agent/user to access local browser session cookies and import account cookies into tooling. Browser/session stores are high-impact credentials, and the artifact does not bound domains, storage, retention, or outputs.
Use `--cookies-from-browser chrome` or configure proxy. ... Requires login. Use Cookie-Editor to import cookies.
Do not grant access to your main browser profile. If needed, use a separate browser profile or limited cookie export, confirm the exact domain/account, and clear or revoke cookies afterward.
Using bot-bypass automation can violate site rules, trigger account or IP restrictions, and create legal or policy risk for the user.
The skill explicitly advertises using Camoufox to bypass WeChat anti-bot protections. That is a material misuse risk even if the user only wants to read an article.
**Read** (Camoufox — bypasses WeChat anti-bot): ... `cd ~/.agent-reach/tools/wechat-article-for-ai && python3 main.py "https://mp.weixin.qq.com/s/ARTICLE_ID"` ... Must use Camoufox.
Avoid anti-bot bypass workflows unless you have explicit authorization. Prefer official APIs, manual viewing, or clearly permitted access methods.
The agent may run tools whose code, version, and installation source are not visible in this skill review.
The supplied package has no install spec or code files, but the instructions depend on many external executables and a local helper path. Those tools are not reviewable from the provided artifact set.
Upstream tools for 13+ platforms. Call them directly. ... `cd ~/.agent-reach/tools/wechat-article-for-ai && python3 main.py "https://mp.weixin.qq.com/s/ARTICLE_ID"`
Install and inspect each required tool separately, pin trusted versions where possible, and do not let the agent execute unknown local helpers automatically.
The skill needs a provider API key for some web/search functions, and that key may be used for requests to the provider.
The SkillBoss API key is expected for the documented web search/scraping provider, but the registry metadata lists no required env vars or primary credential.
requires:
env: [SKILLBOSS_API_KEY] ... headers={"Authorization": f"Bearer {SKILLBOSS_API_KEY}"Use a dedicated, revocable API key and verify the registry metadata or documentation before installing.
If account access is configured, the agent could publish content or images to a social platform when instructed.
The publishing command is user-directed and disclosed, but it can create public content from the user's account.
Use when: ... user asks to post, comment, or interact on supported platforms ... `mcporter call 'xiaohongshu.publish_content(...)'`
Require an explicit final confirmation before any post/comment/publish action and review the exact text, images, account, and destination.
URLs, search terms, and scraped targets may be visible to the external provider.
The skill sends user-supplied URLs or search queries to the SkillBoss/HeyBoss API. This is disclosed and purpose-aligned, but it is a third-party data flow.
`requests.post("https://api.heybossai.com/v1/pilot", ... json={"type": "scraper", "inputs": {"url": "URL"}})`Avoid sending private, internal, or sensitive URLs/queries unless you trust the provider and understand its data handling.