ClawHub Publish

ReviewAudited by ClawScan on May 10, 2026.

Overview

This looks like a legitimate ClawHub publishing helper, but it can rewrite skill files and publish or batch-publish them publicly using the logged-in account without explicit confirmation safeguards.

Use this only if you intentionally want the agent to publish to ClawHub. Before running it, replace the hard-coded paths and profile URL with your own, verify the logged-in ClawHub account, require a diff before any file changes, and approve each publish separately rather than allowing automatic batch publishing.

Findings (6)

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.

What this means

The agent could change a local skill and publish or update it publicly under the logged-in ClawHub account before the user has reviewed the final result.

Why it was flagged

The workflow combines file mutation with public marketplace publishing, and the batch path repeats those actions, but the artifact does not require a final user approval, diff review, or per-skill confirmation.

Skill content
If Korean or personal info is found: ... Rewrite SKILL.md fully in English ... Write with UTF-8 ... clawhub publish ./skills/<skill-name> ... Batch Publish ... run steps 1–4 for each in sequence.
Recommendation

Require the user to select the exact folder, slug, version, and changelog; show a diff before file writes; and ask for explicit confirmation before each publish, especially in batch mode.

What this means

One bad assumption or edit could affect several marketplace listings and make cleanup harder.

Why it was flagged

A translation, sanitization, versioning, or slug mistake could be repeated across multiple public skill publishes without containment.

Skill content
To publish multiple skills at once, run steps 1–4 for each in sequence.
Recommendation

Disable automatic batch publishing by default, process one skill at a time, and require per-skill validation and confirmation.

What this means

Actions will use whichever ClawHub account is currently logged in.

Why it was flagged

The skill relies on the user's authenticated ClawHub session, which is expected for publishing but grants account-level authority to create or update marketplace entries.

Skill content
Logged in: `clawhub whoami` (if not: `clawhub login` via PTY + browser)
Recommendation

Run `clawhub whoami` yourself before publishing and confirm the account is correct.

What this means

Users must trust the npm package and the installed CLI because it performs the actual publish operation.

Why it was flagged

The skill depends on a globally installed npm CLI, which is purpose-aligned but not represented in the install specification or required binaries.

Skill content
clawhub CLI installed: `npm i -g clawhub`
Recommendation

Install only the official ClawHub CLI, verify the package source, and keep the CLI updated.

What this means

Incorrect publish records could be reused later and mislead future planning or automation.

Why it was flagged

The workflow writes publish status into persistent local memory/dashboard files, which may influence later agent context.

Skill content
After successful publish, update both: `C:\MAIBOT\memory\marketplace-strategy.md` ... Obsidian `_DASHBOARD.md`
Recommendation

Review persistent record updates before saving them and keep them limited to the intended workspace.

What this means

A user could mistakenly verify the wrong profile or expect the workflow to fit their environment when it is tailored to someone else's setup.

Why it was flagged

The skill is presented as a general ClawHub publishing helper but contains user-specific paths and a specific profile URL.

Skill content
Target skill folder exists under `C:\MAIBOT\skills\<skill-name>\` ... Then check: https://clawhub.ai/u/jini92
Recommendation

Replace hard-coded paths, suffixes, and profile URLs with user-specific placeholders before using the skill.