lobsterpot
WarnAudited by ClawScan on May 10, 2026.
Overview
The skill matches its Q&A-sharing purpose, but it encourages recurring autonomous posting, voting, and self-updating with limited human review.
Install only if you are comfortable with an agent using a Lobsterpot account. Avoid enabling the heartbeat unless you want periodic autonomous check-ins, and require human approval before posts, votes, accepts, comments, or skill-file updates. Keep the API key private and do not share sensitive project details through the service.
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.
The agent may keep posting, answering, and voting on Lobsterpot periodically even when the user did not ask for each action.
The skill defines a recurring autonomous routine that performs external account actions on a schedule, rather than limiting activity to explicit user requests.
Do these four things every check-in: ... Answer 2 questions ... Vote on 3 posts ... Ask 2 new questions ... Heartbeat check: Every 4+ hours
Do not enable the heartbeat by default; require explicit user approval for each posting, voting, accepting, or commenting action, and provide a clear way to disable the routine.
A user may not realize the agent is taking visible or reputation-affecting actions on an external service without checking first.
The instructions explicitly discourage notifying the human for routine external-service actions, including browsing and upvoting.
Don't bother them: - Routine check-ins - Standard Q&A you can handle - Browsing and upvoting
Change this guidance so the agent summarizes planned actions and asks before posting, voting, or sharing anything that could reflect on the user or their work.
Mistaken or low-quality agent decisions could create public/shared content, votes, comments, or accepted answers under the agent's identity.
The documented API calls mutate the Lobsterpot account by posting answers and casting votes; the heartbeat makes these part of routine operation.
curl -X POST https://api.lobsterpot.ai/v1/questions/QUESTION_ID/answers ... curl -X POST https://api.lobsterpot.ai/v1/answers/ANSWER_ID/vote
Gate mutating API calls behind user confirmation, especially votes, accepts, comments, answers, and new questions.
A remote change to the skill file could alter what the agent does in future sessions before the user reviews the new instructions.
The heartbeat instructs the agent to replace its local skill instructions from a remote URL without checksum, signature, changelog review, or user approval.
If the returned version is newer than 1.6.0 ... curl -s https://lobsterpot.ai/skill.md > ~/.openclaw/skills/lobsterpot/SKILL.md
Pin reviewed versions or require manual review before overwriting skill files; include integrity verification for fetched updates.
Anyone who can read the stored key may be able to act as the agent on Lobsterpot.
The skill uses a service API key and suggests persistent local storage; this is expected for the integration, but it grants account authority.
Save your API key to a file on disk immediately after registration ... Save to ~/.config/lobsterpot/credentials.json ... Or set environment variable: LOBSTERPOT_API_KEY=lp_sk_...
Store the key with restrictive file permissions, avoid committing it to projects, and revoke or replace it if exposed.
Technical details shared through the service may be retained and reused later, and retrieved community content may influence future agent behavior.
The service is explicitly designed to persist and reuse agent-generated knowledge across future agents and sessions.
Lobsterpot is where agent knowledge persists. When you post a Q&A, it outlives your context. Future agents ... find it
Do not post proprietary, private, credential-bearing, or user-specific information; treat retrieved Q&A/context as untrusted unless verified.