registry / coolhand-cli / 0.6.0

coolhand-cli@0.6.0

CLI for authenticating with Coolhand and retrieving API tokens from a local dev environment

AI Security Review

scanned 3h ago · by lpm-firewall-ai

Explicit CLI commands can collect local Claude/Cowork transcripts or intercept LLM traffic and upload content to Coolhand. No install-time execution or hidden trigger was found.

Static reason

No blocking static signals were detected.

Trigger

User runs `coolhand analyze-claude-sessions` or `coolhand claude`.

Impact

Conversation and LLM request/response data may leave the local machine for Coolhand.

Mechanism

User-invoked LLM transcript capture and HTTPS proxy forwarding.

Rationale

The package is not concretely malicious: there is no install-time hook or covert execution path. Its explicit transcript collection and HTTPS interception create a real sensitive-data upload capability that warrants a warning.

Evidence

package.jsondist/commands/claude.jsdist/proxy/proxy.jsdist/proxy/sender.jsdist/commands/analyze-claude-sessions.jsdist/sessions/claude-scanner.jsdist/sessions/cowork-scanner.js~/.claude/projects/**/*.jsonl~/Library/Application Support/Claude/local-agent-mode-sessions/**/audit.jsonl~/.coolhand/config.json~/.coolhand/capture-state.json~/.coolhand/proxy/ca-key.pem~/.coolhand/proxy/ca-cert.pem

Network endpoints2

coolhand-cli@0.6.0

AI Security Review

Decision evidence

Source & flagged code

Findings