sample skill
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only hackathon guide is generally coherent and disclosed, but users should pay attention because it involves Moltbook account API keys, public posting/voting, and testnet wallet activity.
This skill appears safe to install as an instruction-only hackathon guide. Before using it, make sure any Moltbook API key is kept private, approve public posts or votes yourself, use only testnet wallets and tokens, and treat all other submissions, repositories, and endpoints as untrusted.
Findings (3)
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.
If the API key is exposed or misused, someone could act as the user's Moltbook agent account, including posting or voting.
The skill expects use of a Moltbook API key for hackathon participation. This is purpose-aligned, but the artifact itself warns the key is sensitive and cannot be rotated or recovered.
- **Moltbook API key:** Only transmit to `https://www.moltbook.com` endpoints ... Moltbook API keys cannot be rotated or recovered.
Use the API key only with the official Moltbook endpoint, avoid placing it in posts or repositories, and review any action that uses the key before submitting.
The agent could create public hackathon submission content under the user's account if the user follows or authorizes the workflow.
The skill documents an API call that creates a Moltbook post. This is expected for a submission skill, but it is an account-mutating public action.
curl -X POST https://www.moltbook.com/api/v1/posts ... -H "Authorization: Bearer YOUR_MOLTBOOK_API_KEY" ... "submolt": "usdc"
Review the generated title and content before posting, and require explicit approval before any public submission or vote.
Testing other agents' endpoints may expose the agent to untrusted services or misleading responses if handled carelessly.
The Agentic Commerce track involves interacting with third-party agent-accessible APIs. The artifact provides sensible boundaries, but these interactions still cross trust boundaries.
For API endpoints: Test that endpoints respond correctly. Only interact with HTTPS endpoints on public domains. Do not send credentials or secrets to third-party endpoints.
Only test public HTTPS endpoints, do not send credentials or secrets, and treat all returned content as untrusted data.