psyclaw@0.29.17
A standalone, evidence-grounded social-science research agent with research contracts, evidence gates, and recoverable workflows
AI Security Review
scanned 2mo ago · by lpm-firewall-aiSource inspection found no install-time execution, self-dependency chain, credential exfiltration, or unconsented foreign agent-control-surface mutation. The flagged files implement constrained runtime features or test fixtures, not a concrete attack.
Decision evidence
public snapshot- The manifest has no preinstall, install, or postinstall hook; its only lifecycle hook is prepublishOnly.
- The flagged MCP transport is runtime code for a configured command, uses argument-vector spawning without a shell, and does not inherit the parent environment.
- The flagged vendor test file is ordinary pytest fixture code and contains no payload execution.
- The flagged TUI install path requires an explicit Y confirmation before calling the installer.
- The package provides user-invoked agent, plugin, and update functions that can install software or write project-local PsyClaw configuration.
- It ships a Windows fd.exe helper binary, but no source evidence tied it to automatic execution or malicious behavior.
Source & flagged code
6 flaggedPackage source references dynamic require/import behavior.
dist/src/network-routing.jsView on unpkg · L170Findings
Affected versions and remediation
This report applies to psyclaw@0.29.17.
See all versions of psyclaw for other recorded verdicts.
- Review the evidence and your use of psyclaw@0.29.17 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.