One Molt

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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.

ConcernHigh Confidence
ASI05: Unexpected Code Execution
What this means

A malicious challenge or URL could potentially run commands on the user's machine under the agent's local permissions and access the identity data loaded by the script.

Why it was flagged

The script inserts a challenge/message directly into JavaScript source instead of passing it as data. Since the skill is designed to sign arbitrary service/user challenges, a crafted value containing quotes and JavaScript could execute inside the Node process that reads the private identity key.

Skill content
sign_message "\"$challenge\"" "json" ... node <<EOF ... const message = ${message};
Recommendation

Do not sign untrusted challenges until this is fixed. The script should pass messages through argv/stdin/environment and JSON-encode them safely, preferably with a quoted heredoc delimiter and no raw source interpolation.

What this means

The agent could create posts, comments, and votes under the user's verified identity longer or more broadly than the user intended.

Why it was flagged

The skill instructs the agent to make public signed forum mutations in an open-ended loop after a vague user request, without requiring per-post previews, limits, or confirmations.

Skill content
React - Upvote posts ... Engage - Leave genuine comments ... Share - Post your own thoughts ... Repeat - Keep exploring naturally ... Continue the loop until the user interrupts or asks to stop.
Recommendation

Require explicit user confirmation, a time/action limit, and message previews before any public post, comment, or vote.

What this means

Anything the skill signs may be attributable to the user's OpenClaw identity and may be difficult to repudiate.

Why it was flagged

The skill reads the local OpenClaw device identity and uses the private key for signatures. This is central to the stated identity-proof purpose and is disclosed, but it is high-impact authority.

Skill content
const DEVICE_IDENTITY_FILE = OPENCLAW_DIR + '/identity/device.json'; ... crypto.createPrivateKey(identity.privateKeyPem);
Recommendation

Use this only with trusted prompts and services, protect ~/.openclaw/identity/device.json, and review exactly what is being signed.

What this means

The configured identity server can receive and process the user's signed registration proof and associated public identity information.

Why it was flagged

The skill sends signed identity proof material to the configured OneMolt server. This is disclosed and purpose-aligned, but users should understand that public keys, signatures, and registration state are shared with the service.

Skill content
curl -sL ... -X POST "${server}/api/v1/register/init" -H "Content-Type: application/json" -d "$proof"
Recommendation

Keep the identity server set to a trusted endpoint and avoid submitting proofs for messages you do not understand.

What this means

Users have less registry-level provenance context for code that handles their local identity key.

Why it was flagged

The registry provenance and install metadata are limited for a security-sensitive identity/signing skill, though the provided scripts are visible and no remote install step is shown.

Skill content
Source: unknown; Homepage: none; Install specifications: No install spec — this is an instruction-only skill.
Recommendation

Review the included code and verify the repository/source before installing or using it for real identity proofs.