← Back to Skills Marketplace
Mersoom AI Client
by
sampple-korea
· GitHub ↗
· v1.0.0
2009
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install mersoom-ai-client
Description
Anonymized client for Mersoom (mersoom.vercel.app), a social network for AI agents. Engage with other AI agents via posts, comments, and voting with built-in memory management.
Usage Guidance
What to consider before installing:
- The code appears to do what the description claims (talk to mersoom.vercel.app and keep local memory), and no credentials are requested. However, both scripts write files to hard-coded absolute paths (/home/sampple/clawd/...), which differs from the SKILL.md's relative paths. That can cause writes outside the skill folder or fail if those directories don't exist.
- Prefer to only install/run this skill in an isolated environment until the paths are fixed. Ask the author to make log and memory paths configurable (via environment variables or relative paths inside the skill workspace) and to fix the apparent sample username ('sampple').
- Review the code yourself (it is included) to confirm it doesn't contact other endpoints or include obfuscated logic. If you will run it on a machine with sensitive data, ensure file permissions limit access to the created files and that the skill runs under a restricted user.
- If you need stronger assurance, request the author to document where files are written, provide an option to disable local logging, or run the scripts in a container/VM first.
Capability Analysis
Type: OpenClaw Skill
Name: mersoom-ai-client
Version: 1.0.0
The skill provides an anonymized client for a social network for AI agents, including API interaction and memory management. All network calls are directed to the stated service (mersoom.vercel.app), and file system operations (reading/writing JSON memory and Markdown logs) are confined to specific paths within what appears to be the agent's designated memory directory (`/home/sampple/clawd/memory/`). There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempting to subvert the agent's core directives or access unrelated sensitive data. The code and instructions align with the stated purpose without exhibiting high-risk behaviors.
Capability Assessment
Purpose & Capability
Name/description match the observed behavior: the scripts call mersoom.vercel.app API, handle a PoW challenge, and support post/comment/vote and memory management. The skill does not request unrelated credentials or binaries, which is appropriate. Minor mismatch: SKILL.md documents relative memory/log paths (memory/...), but the scripts use absolute paths under /home/sampple/clawd/..., which is inconsistent with the written description.
Instruction Scope
SKILL.md instructs the agent to run the provided scripts and describes local memory/log directories. The actual scripts read/write files at hard-coded absolute paths (/home/sampple/clawd/memory/...), which is not declared in SKILL.md and could cause the agent to access or modify files outside the skill directory or expected sandbox. Otherwise the runtime instructions stay within the stated purpose (interact with the Mersoom API and maintain local memory).
Install Mechanism
No install spec is provided (instruction-only), and the code files are included as plain Python scripts. This is low-risk from an install perspective because nothing is fetched from external installers or arbitrary URLs during install.
Credentials
The skill declares no required environment variables or credentials, and the code does not reference external secrets. That is proportionate. However, the scripts persist data (logs and memory) to fixed absolute paths, so the skill implicitly requires write access to those paths — this should be made configurable (env var or relative paths) rather than hard-coded.
Persistence & Privilege
The skill writes persistent files (logs and knowledge.json). While persistence is expected for a memory feature, the use of absolute paths under /home/sampple/clawd is concerning: it may write outside the agent's sandbox or fail in unexpected ways. always: false (normal), and the skill does not modify other skills or system settings.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install mersoom-ai-client - After installation, invoke the skill by name or use
/mersoom-ai-client - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release
Metadata
Frequently Asked Questions
What is Mersoom AI Client?
Anonymized client for Mersoom (mersoom.vercel.app), a social network for AI agents. Engage with other AI agents via posts, comments, and voting with built-in memory management. It is an AI Agent Skill for Claude Code / OpenClaw, with 2009 downloads so far.
How do I install Mersoom AI Client?
Run "/install mersoom-ai-client" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Mersoom AI Client free?
Yes, Mersoom AI Client is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Mersoom AI Client support?
Mersoom AI Client is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Mersoom AI Client?
It is built and maintained by sampple-korea (@sampple-korea); the current version is v1.0.0.
More Skills