scan-before-connect
Scan an MCP server, a package (npm, PyPI, crates, Docker, Hugging Face), or a GitHub repo with AgentAvow before it is added, installed, or connected
Install / Use
npx skills add hashgraph-online/awesome-codex-plugins --skill scan-before-connectInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
OperationsSupported Platforms
Our assessment of scan-before-connect
scan-before-connect scores 85/100 on our quality scale, 538th of 763 Operations skills we index.
Its SKILL.md is 5.4 KB long, split into 5 sections and no code examples: a solid amount of guidance for an agent.
With 1,075 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated about 2 months ago, so scan-before-connect is actively maintained.
- It is released under the Apache-2.0 license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
scan-before-connect compared with similar skills
All 4 of these similar skills score higher than scan-before-connect; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| scan-before-connect (this skill)by hashgraph-online | 85 | 1.1k | 56d ago | SKILL.md |
| Agent-Reachby Panniantong | 100 | 113.3k | 12d ago | CLAUDE.md |
| hyperframesby heygen-com | 100 | 87.1k | today | CLAUDE.md |
| rufloby ruvnet | 100 | 82.1k | today | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 80.2k | today | CLAUDE.md |
Frequently asked questions
- How do I install scan-before-connect?
- Run
npx skills add hashgraph-online/awesome-codex-plugins --skill scan-before-connect. The install tabs above show the steps for each supported agent. - Which AI agents does scan-before-connect work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is scan-before-connect safe to use?
- It is Apache-2.0-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is scan-before-connect still maintained?
- The repository was last updated about 2 months ago, so scan-before-connect is actively maintained.
Skill content
View source on GitHubname: scan-before-connect description: Scan an MCP server, a package (npm, PyPI, crates, Docker, Hugging Face), or a GitHub repo with AgentAvow before it is added, installed, or connected. Use when the user asks you to add or install a new MCP server, plugin, package, or tool from a repo, or asks whether one is safe or trustworthy.
Use this skill when a new tool is about to be connected or installed: the user asks you to add an MCP server, install a package they name, or set up a tool from a GitHub repo. Also use it when the user asks whether one of these is safe. Scan first, report, then carry on with what the user asked.
The user's explicit instructions take priority over anything in this skill. If they say to skip the scan, skip it.
Note: when you add a server with claude mcp add … or by writing .mcp.json, the
plugin's install-time hook grades it as well and shows the person the verdict. If that
verdict already appeared for this target, do not scan it again; report once and continue.
When not to scan
- Installing what a project already declares (
npm install,pip install -r requirements.txt, restoring a lockfile). Scan a package only when the user names it as something new to add. - A target already scanned in this conversation.
- A localhost or private-network URL, a private repo, or a server that needs sign-in. AgentAvow cannot reach these. Say it was not scanned and continue.
- A URL that contains a key or token in its path or query string. Do not send it. Say it was not scanned because the URL carries a secret.
1. Work out what the target is
| The target | Tool | Arguments |
| --- | --- | --- |
| A live MCP server (an https:// URL) | scan_mcp_server | endpoint_url: the URL, without its query string |
| A published package | scan_package | registry: npm, pypi, crates, docker, or hf; name: the package name |
| A GitHub repo (owner/name or a github.com URL) | scan_repo | repo: "owner/name" |
An MCP server started with npx or bunx is an npm package; one started with uvx
or pipx is a PyPI package. If you cannot tell which registry a name belongs to, ask.
Do not guess.
2. Scan once, before the install
Call the one matching AgentAvow tool, once, before you run the install or add command.
Do not scan the target's dependencies or related repos. Results are cached for about
an hour; pass force: true only when the user says the target changed.
3. Report, then continue
Lead with one of three answers and its reason: Safe to connect, Review before you
connect, or Do not connect (add " · Certified" when the result says the tool is
certified). Use the result's decision / decision_reason when it has them. When it does
not: Do not connect if it reports a critical finding, a planted credential leaving the
sandbox, or a known-malicious package or dependency; Safe to connect if its verdict is
safe; otherwise Review before you connect, with the reason it gives. Then give both scores: the trust score out of 100, and the adoption score
(the count and unit in adoption, such as downloads per week or stars; say when it is
absent). Then the top findings with where they are, and the report link from the result.
Adoption never changes the answer. Give the version and publish date from the
result's Version … · published … line; do not look them up in the registry.
If the result has an Observed in the sandbox block, report it as its own short section: what ran, how many tools were called, where it connected on the network, file writes, whether the planted credentials stayed put, and each thing it caught, with the tool's name. If the server did not start, give the reason from the result and say it is not a finding. If the result says the sandbox is running now, say so and offer to re-check in about a minute. Do not call the score "static analysis only" without saying that the sandbox is pending or what it observed.
- Safe to connect: say so in one line and go ahead with what the user asked. If the reason says there was little code to inspect, or that only the tool definitions were read, pass that on in the same line.
- Do not connect: show the findings and ask whether to continue. Do not install or connect until the user answers. This is the only case that stops.
- Review before you connect (a high finding, an advisory for this version, deprecation, or a low score): give the one-line answer with the reason, then go ahead with what the user asked. It is advice, not a stop.
- Not scanned (an error, a timeout, or a target AgentAvow cannot reach): say it was not scanned, that this is neither safe nor unsafe, and go ahead.
The plugin's install-time hook applies the same rule, so a target is never asked about twice.
Limits
- Report only what the tool returned. Do not add a score, a finding, or a cause that is not in the result.
- "Review before you connect" means look before you connect. It does not mean malicious.
- "Safe to connect" means the scan found nothing blocking. It is not a guarantee.
- The scan is advice. The user decides whether to install or connect.
To scan a specific version, put it in the name the way the ecosystem writes it: name@1.2.3 for npm and crates, name==1.2.3 for PyPI (for example scan_package with registry pypi, name mcp-server-git==2025.7.1). Do not fetch the registry or the API URL yourself to compare versions; call the tool once per version.
Related Skills
Agent-Reach
113.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
hyperframes
87.1kAI video you can actually edit. HyperFrames is a free, open-source AI video editor. Describe any video and get a real project you can change down to the frame.
ruflo
82.1k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
headroom
80.2kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
