@companion-ai/feynman@0.3.27
Research-first CLI agent built on Pi and alphaXiv
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThis is not malicious under the stated boundary because the risky setup occurs after explicit CLI execution and is scoped to Feynman/Pi runtime setup. Warn due to automatic runtime patching and optional system package installation.
Decision evidence
public snapshot- CLI imports a runtime patcher before launching the app.
- Patcher deletes and extracts its bundled runtime workspace.
- Patcher rewrites configured Pi subagent package files.
- Global macOS CLI runs Homebrew to install Pandoc when absent.
- The preinstall hook only enforces a Node-version range.
- Archive extraction verifies a packaged SHA-256 first.
- No credential harvesting, covert exfiltration, or install-time foreign-agent mutation was confirmed.
Source & flagged code
14 flaggedPackage defines install-time lifecycle scripts.
package.jsonView on unpkgPackage source references a known benign dynamic code generation pattern.
dist/workbench-web/assets/3Dmol-D4OPhBxL.jsView on unpkg · L1692Findings
Affected versions and remediation
This report applies to @companion-ai/feynman@0.3.27.
See all versions of @companion-ai/feynman for other recorded verdicts.
- Review the evidence and your use of @companion-ai/feynman@0.3.27 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.