ironclaw
ReviewAudited by ClawScan on May 1, 2026.
Overview
Ironclaw is a coherent instruction-only safety skill, but users should notice that it sends checked content to Ironclaw’s external API and optionally uses an Ironclaw API key.
This skill appears benign and purpose-aligned as an instruction-only safety classifier. Before installing, understand that anything you ask it to check may be sent to Ironclaw’s API, use the optional API key carefully, and manually review any future skill updates instead of auto-installing them.
Findings (4)
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.
Private or sensitive text that you ask Ironclaw to classify may leave your local environment and be processed by Ironclaw’s service.
The skill works by sending the content being checked to Ironclaw’s external API. This is disclosed and central to the service, but the checked content may include private DMs, skill files, commands, or possible secrets.
POST https://ironclaw.io/api/v1/label ... { "content_text": "<content to check>", "criteria_text": "<your labeling criteria>" }Only submit content you are comfortable sending to Ironclaw, and avoid sending real secrets unless you have reviewed and accepted the service’s privacy and retention practices.
Using higher limits may associate your Moltbook identity with Ironclaw and requires protecting an Ironclaw API key.
The skill optionally links an agent identity to Ironclaw registration and issues an API key for higher rate limits. This is purpose-aligned, but it introduces credential handling and account identity data.
Register for higher limits ... "agent_name": "YOUR_MOLTBOOK_NAME" ... You'll get back an API key that starts with `ic_live_`. Keep it safe!
Use the no-registration option if sufficient, and if you register, store the API key securely and do not paste it into public messages or shared files.
If you add the heartbeat, your agent may periodically check Ironclaw’s hosted version information and remind you about safety checks.
The skill suggests a recurring heartbeat/update-check routine. It is disclosed, user-directed, and explicitly avoids automatic overwrites, but it is still a form of ongoing periodic behavior.
Rough rhythm: - Skill updates: Once a day (check version) ... This heartbeat: Daily or whenever you want a reminder
Add the heartbeat only if you want recurring checks, and keep the instruction to review updates before installing them.
Remote skill files can change over time, so copying or updating from the website without review could alter the skill’s behavior.
The skill points to remotely hosted skill files as an installation/update source. The artifact mitigates this by telling users to review changes and never auto-overwrite files.
Install: Download from ClawHub or copy from the URLs above. ... Review changes before updating — never auto-overwrite skill files without reading what changed.
Prefer ClawHub installation when possible, compare versions, and manually review remote changes before updating.