negotium@0.22.5
⚠ Under reviewInstall the Negotium multi-agent runtime and CLI with one package
AI Security Review
scanned 19h ago · by lpm-firewall-aiThe inspected lifecycle installers and flagged agent integration locations show no confirmed attack behavior. Manual review is warranted because the declared @anthropic-ai/sdk dependency’s source was unavailable, so its install behavior could not be verified.
Decision evidence
public snapshot- package.json declares @anthropic-ai/sdk, and the Claude provider imports its message types; the dependency source was unavailable in the supplied snapshot.
- The inspected agent integration code resolves local Codex authentication paths, handles Claude session files during session operations, and creates a private Codex hook bridge for a provider turn.
- The postinstall hook installs versioned browser and shell binaries into Negotium’s state directory and checks each download against a pinned hash.
- The inspected integration locations show no confirmed attack behavior; their observed file operations support agent sessions and a private per-turn hook.
- The postinstall installers target Negotium’s own binary directory and verify downloaded binaries against pinned hashes.
Source & flagged code
20 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgPackage source references child process execution.
dist/hosted-agent.jsView on unpkg · L50Findings
Affected versions and remediation
This report applies to negotium@0.22.5.
See all versions of negotium for other recorded verdicts.
- Review the evidence and your use of negotium@0.22.5 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.