registry / @kahinmcp/kahin / 0.3.2

@kahinmcp/kahin@0.3.2

Kahin — CDP ansiklopedisi + anti-detect browser otomasyon MCP sunucusu. Kurulumda 22 AI CLI aracını otomatik tespit edip kendini kaydeder. Obscura ve Camoufox motorlarını kullanır, motor seçimini ajana bırakır.

AI Security Review

scanned 10d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the package automatically modifies detected configurations for numerous unrelated AI-agent clients. It registers kahin as an MCP server, causing those clients to invoke the package later.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm postinstall of @kahinmcp/kahin@0.3.2

Impact

Foreign AI clients can be silently configured to launch the package's MCP server.

Mechanism

unconsented broad AI-agent MCP configuration injection

Policy narrative

At npm postinstall, bin/setup.mjs scans for existing configurations belonging to 22 AI CLI/editor clients and adds a kahin MCP entry to every configuration it finds. This is not limited to a package-owned integration or an explicit setup command. The injected command later starts the bundled Python MCP server through bin/kahin.mjs.

Rationale

The postinstall hook performs unconsented mutation of a broad, foreign AI-agent control surface. This meets the blocking policy regardless of the package's stated browser-automation purpose.

Evidence

package.jsonbin/setup.mjsbin/kahin.mjslib/kahin-0.3.2-py3-none-any.whl~/.claude.json~/.codex/config.toml~/.config/Code/User/mcp.json~/.config/opencode/opencode.json

@kahinmcp/kahin@0.3.2

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code