oh-my-opencode@4.16.3
The Best AI Agent Harness - Batteries-Included OpenCode Plugin with Multi-Model Orchestration, Parallel Background Agents, and Crafted LSP/AST Tools
AI Security Review
scanned 3mo ago · by lpm-firewall-aiSource inspection contradicts the scanner's malicious classification, but the bundled agent plugin has real lifecycle, telemetry, configuration-migration, and executable-provisioning capabilities. Per policy this is a warn-level first-party agent extension risk, not malware.
Decision evidence
public snapshot- package.json runs postinstall.mjs on npm install.
- postinstall.mjs removes matching OpenCode plugin-cache directories.
- Codex SessionStart hook runs auto-update.mjs, which migrates OMO/Codex configuration.
- codegraph/dist/cli.js downloads checksum-pinned executables and later provisions them under ~/.omo/codegraph.
- telemetry hook sends activity through PostHog to us.i.posthog.com.
- postinstall.mjs only checks a local CLI, resolves its platform package, and invalidates this package's named cache entries.
- No postinstall network request, credential harvesting, or foreign agent-config write was found.
- Codegraph download URLs and SHA-256 values are fixed in source.
- The PowerShell file is an explicit ast-grep smoke test, not an inert hidden payload.
Source & flagged code
14 flaggedPackage 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.
bin/oh-my-opencode.jsView on unpkg · L4Findings
Affected versions and remediation
This report applies to oh-my-opencode@4.16.3.
See all versions of oh-my-opencode for other recorded verdicts.
- Review the evidence and your use of oh-my-opencode@4.16.3 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.