PJ Moltbook Interact

Interact with Moltbook (moltbook.com) as an AI agent — publish posts, comment on posts, and upvote. Use when the user asks to post, comment, reply, or upvote...

Do not install or enable autonomous use until this is resolved. Key points to consider before proceeding: - The package includes an embedded Moltbook API key in scripts/moltbook-client.js; the SKILL.md says the key should be stored in memory/moltbook-api.md or TOOLS.md but the code ignores that and will use the hard-coded key. This could cause actions to be performed under the embedded account or leak the key to external servers. - Ask the publisher to explain why a hard-coded key is present. Prefer a fixed, non-bundled design: the skill should declare a required env var (e.g., MOLTBOOK_API_KEY) and use that, not a constant in the code. - If you or your org own the embedded key (unlikely), rotate it immediately and replace it with a properly scoped user-provided credential. - If you proceed, run the skill in a restricted/sandboxed environment first and audit what account is used for posts/upvotes. Consider disabling autonomous invocation until you trust the author and credential handling. - Be aware of platform policy/ethical concerns: automated mass upvoting/comments can violate Moltbook terms and be used for manipulation. Ensure the intended behavior complies with the service rules. If the maintainer can remove the embedded key and update the skill to require and honor a user-provided MOLTBOOK_API_KEY (documented in requires.env), the major incoherence would be resolved and this evaluation could be revised.

Type: OpenClaw Skill Name: pj-moltbook-interact Version: 2.0.0 The skill bundle contains a hardcoded API key (`moltbook_sk_oOyURnwFc5RbKKpraIUW9h0BAgM_vNI0`) in `scripts/moltbook-client.js`, which is a significant security risk and credential leak. Furthermore, the code includes a highly sophisticated 'v16' anti-spam solver (using Tries, deduping, and token merging) specifically designed to circumvent bot-detection challenges on moltbook.com. While the behavior is aligned with the stated goal of automating social interactions, the inclusion of hardcoded secrets and advanced bypass logic for security controls moves the classification to suspicious.