Guide · Agent skill
How to use the Iron Proxy Gateway agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The iron-proxy-gateway skill teaches an AI agent how to work inside NanoClaw's Iron Proxy, a gateway that sits between the agent and the internet. It is aimed at developers and operators who run NanoClaw agents that need to call GitHub or other external APIs without being handed raw tokens. After this guide you will know what the skill changes, how to install it, how to ask for things in plain language, and how to read a blocked request.
Iron Proxy Gatewaynanocoai/nanoclaw · MIT Download the skill Quick start
- 1Download and unzip. Download the iron-proxy-gateway ZIP from this page and unzip it.
- 2Place the folder. Move the folder into ~/.claude/skills/ or your project's .claude/skills/, or upload it under Settings → Capabilities → Skills.
- 3Confirm the environment. Check that NanoClaw runs with Iron Proxy, HTTP_PROXY and HTTPS_PROXY are set, and the Iron Proxy CA is injected.
- 4Ask for the service. Describe the task and the service, for example a gh request against GitHub.
- 5Run the connect command. The agent runs ncl groups connect --host with the API hostname and shows the connect_url.
- 6Complete setup in the operator console. The operator configures the credential, grant, and destination in the Iron Proxy operator console.
- 7Verify. Once setup is done, the agent runs a credentialed request to confirm access before reporting success.
What is an agent skill, and what does this one do?
An agent skill is a folder with a SKILL.md file that gives an AI agent extra instructions for a specific task. The agent loads these instructions when a request matches the skill's description.
With this skill, the agent's outbound HTTP and HTTPS requests pass through the session's Iron Proxy. When a request needs a credential, the proxy waits for human approval before inserting it. The agent only ever sees a placeholder, never the real token.
- Routes external API calls, including GitHub through gh, via the proxy
- Points the user to the operator console to configure credentials and grants
- Explains blocked requests such as HTTP 403 without guessing the cause
- Avoids adding MCP servers or local logins just to connect an account
Install the skill from the ZIP download
Download the ZIP for iron-proxy-gateway from this page and unzip it. Place the resulting folder in one of two locations, depending on how you want it available:
Personal use across projects: ~/.claude/skills/. Project-only use: the .claude/skills/ folder inside your project. The skill folder should contain SKILL.md and instructions.md.
If you use the Claude apps instead, upload the skill under Settings → Capabilities → Skills.
- Option A: unzip into ~/.claude/skills/
- Option B: unzip into a project's .claude/skills/
- Option C: upload through Settings → Capabilities → Skills in the Claude apps
Check the environment before you rely on it
The skill only works inside a NanoClaw runtime with Iron Proxy enabled. The skill's compatibility note says it requires HTTP_PROXY, HTTPS_PROXY, and the Iron Proxy CA injected by NanoClaw. Without these, the instructions will not match what the agent sees.
Connecting an account also uses the ncl command-line tool. The operator, not the agent, configures credentials and grants in the Iron Proxy operator console.
- NanoClaw runtime with Iron Proxy enabled
- HTTP_PROXY and HTTPS_PROXY environment variables set
- Iron Proxy CA certificate injected by NanoClaw
- The ncl command-line tool available
- Operator access to the Iron Proxy operator console
- External accounts such as GitHub, configured by the operator
Ask for account access in natural language
You do not need a special command to trigger the skill. Describe the task and name the service. For example, ask the agent to check open issues in a repository with gh, or to call a specific API. When the request involves an external account, the agent should run the shared connect command for that API hostname.
The command used by the skill is ncl groups connect with a --host flag set to the API hostname. The agent returns the exact connect_url and explains its action. An operator_console handoff means the operator must finish setup in the gateway. It is not an OAuth login link, and running the command does not grant access.
- Example request: "Use gh to list open pull requests on this repo."
- Example request: "Connect this agent to the GitHub API."
- Example request: "My API call is returning 403. Why is it blocked?"
ncl groups connect --host <API hostname>Read blocked requests and approvals correctly
A bare 403 does not show which layer rejected the request. It could be the destination rule, the credential grant, human approval, or the upstream API. The skill tells the agent to report the hostname and the observed error and to ask for a host-side check, rather than assuming credentials were never injected.
A 401 is not proof that injection failed either, since the stored token may be invalid. Request approval and account connection are separate steps. Nothing is treated as granted until the operator completes configuration and a credentialed request succeeds.
- Report the hostname and exact error you observed
- Treat a pending approval as not yet granted
- Verify with a credentialed request before reporting success
Safety notes: scripts and network access
This skill runs scripts and uses the network. Its job is to make outbound calls through the proxy, so it should be used only in environments where that traffic is expected and governed.
The skill sets firm rules. The agent must never ask for, print, or store a raw API key or OAuth token. It must not bypass the configured proxy or its CA validation, and it must not weaken TLS when proxy errors appear. Those errors are treated as policy or configuration problems to hand to the operator.
- Never paste or store raw API keys or OAuth tokens
- Never bypass the proxy or CA validation
- Gateway-managed placeholders such as GH_TOKEN=gateway-managed are the only non-secret values allowed in clients
What people use it for
- Letting an agent list issues or pull requests on GitHub through gh without seeing the token
- Diagnosing why an agent's outbound API call returns 403 and which layer is blocking it
- Preparing a request for an operator to set up a new destination rule for a partner API
- Running a credentialed API check after the operator finishes configuring access
Frequently asked questions
How do I install the iron-proxy-gateway skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use or a project's .claude/skills/ folder for one project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
What do I need to use it?
You need a NanoClaw runtime with Iron Proxy enabled, the HTTP_PROXY and HTTPS_PROXY environment variables, the Iron Proxy CA certificate injected by NanoClaw, the ncl command-line tool, and operator access to the Iron Proxy operator console. External accounts such as GitHub must be configured by the operator.
Is the iron-proxy-gateway skill free?
The skill is published under the MIT licence. The source does not mention pricing for it, so check the repository for licence terms.
Does this skill log me into GitHub with OAuth?
No. The connection is handled through the Iron Proxy operator console, not an OAuth link. The connect command shows a connect_url and an action, and an operator_console handoff means the operator must configure the credential and grant.
Why does my agent get a 403 error?
A 403 alone does not show which layer blocked the request. It could be the destination rule, the credential grant, a pending human approval, or the upstream API. Ask the operator for a host-side check and share the hostname and exact error.
Can the agent use my GitHub token directly?
The skill forbids it. The agent should never receive or store a raw token. The proxy inserts the credential after approval, and the client only uses a non-secret placeholder such as gateway-managed for gh.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you run NanoClaw agents with Iron Proxy and want GitHub or other APIs called without handing the agent raw tokens. Skip it if you only need general coding help without external accounts, or if your setup lacks the Iron Proxy environment, since the instructions depend on it.
Download the skill