About Slopspotter
Slopspotter is an MCP server in the Social category: detects AI-hallucinated (slopsquatted) Python packages before they get installed. CLI + stdio MCP server for AI coding agents.
Install
git clone https://github.com/conranpearce/slopspotterThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Slopspotter
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
slopspotter
Detects AI-hallucinated Python packages before they get installed.
Checks a package against a curated hallucinated packages database and verifies its existence on PyPI in real time, returning a risk assessment with a clear install recommendation.
Installs two tools:
slopspotter— CLI for manual package checksslopspotter-mcp— stdio MCP server for AI coding agents
Installation
pip install slopspotter
Usage
slopspotter requests
{
"package": "requests",
"exists": true,
"pypi_unavailable": false,
"in_slop_db": false,
"risk_level": "low",
"safe_to_install": true,
"recommendation": "PROCEED: 'requests' appears safe to install."
}
Example high risk output:
{
"package": "cohere-ai-sdk",
"exists": false,
"pypi_unavailable": false,
"in_slop_db": true,
"risk_level": "high",
"safe_to_install": false,
"recommendation": "BLOCK: 'cohere-ai-sdk' does not exist on PyPI and is flagged as a hallucinated package."
}
Exit codes: 0 = low risk, 1 = medium, 2 = high, 3 = critical.
Risk Levels
| Level | Meaning |
|---|---|
low |
Exists on PyPI, not flagged |
medium |
Flagged with low historical risk |
high |
Not on PyPI (hallucinated), flagged, or PyPI unreachable |
critical |
Flagged and actively installable on PyPI |
How It Works
Each package is checked against a curated database of known AI-hallucinated packages and verified against PyPI in real time. The database is maintained by the project author.
MCP Server
slopspotter ships with a stdio MCP server for use with AI coding agents (Claude Code, Cursor, etc.). Once connected, the agent will automatically call assess_package before any install and block unsafe packages.
Claude Code
claude mcp add slopspotter uvx --from slopspotter slopspotter-mcp
Or add manually to your project's .mcp.json:
{
"mcpServers": {
"slopspotter": {
"type": "stdio",
"command": "uvx",
"args": ["--from", "slopspotter", "slopspotter-mcp"]
}
}
}
If you already have slopspotter installed via pip:
{
"mcpServers": {
"slopspotter": {
"type": "stdio",
"command": "slopspotter-mcp"
}
}
}
Other agents (Cursor, Windsurf, etc.)
Use the same JSON config — most agents support the mcpServers format.
Usage
Once connected, your agent will automatically assess packages before installing. You can also ask directly:
"Is
requestssafe to install?"
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.