Gemini Spark Core
SuspiciousAudited by ClawScan on May 10, 2026.
Overview
The skill mostly does what it says, but it includes a specific-looking Moltbook API key/agent identity and can post publicly, so users should review it before use.
Do not use the embedded sample key. Install only after verifying the publisher/source, replacing the credential with your own Moltbook API key, and configuring the agent to ask before posting or replying publicly.
Findings (5)
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.
A user could accidentally use or expose someone else's API key, post under the wrong Moltbook identity, or trust a leaked credential.
The skill documentation embeds a concrete credential-like Moltbook API key and agent identity, while the integration can create posts and comments using a bearer token.
"api_key": "clh_-Y5CvhWaIDPHXS3AIn...", "agent_name": "Gemini-Spark"
Remove and rotate the embedded key, replace it with a placeholder, declare the credential requirement, and configure only the user's own Moltbook API key.
Users may be falsely reassured about credential handling and data leaving the machine.
These safety claims are misleading in context because SKILL.md includes a credential-like API key and the script sends API requests to Moltbook.
"No credentials in repo" and "Local only - All processing happens on your machine"
Correct the security claims, clearly state that Moltbook requests are sent to the external API, and ensure no real credentials are present in the package.
If invoked carelessly, the agent could publish unwanted public content.
The helper script can publish replies and new posts to the user's Moltbook account; this is central to the skill but is a meaningful mutation authority.
api_call POST "/posts/${post_id}/comments" ... and api_call POST "/posts" ...Use create/reply only after explicit user approval, preferably with a preview of the exact post or reply text.
A stale or tampered reply log could cause the agent to skip or repeat engagements.
The skill suggests maintaining persistent state that can affect future engagement decisions.
Log file: `/workspace/memory/moltbook-replies.txt` ... Check post IDs against existing replies before posting
Keep the log scoped to this skill, make it reviewable, and avoid storing sensitive post content unless needed.
Users may have difficulty verifying they are installing the intended Moltbook skill before trusting it with credentials.
The package provenance is under-specified, and the evaluated slug differs from SKILL.md's `moltbook` name and the origin file's `moltbook-interact` slug.
Source: unknown; Homepage: none; Slug: gemini-spark-core
Verify the package source and publisher, and align the registry slug, skill name, origin metadata, and repository links.