@whalent/agent-core@0.3.504
⚠ Under reviewCore runtime for Whalent Agent
AI Security Review
scanned 10d ago · by lpm-firewall-aiThe package has no install hook, but its manifest selects a very large obfuscated entrypoint whose runtime behavior I could not resolve through static inspection. The visible skill-install helpers leave their callers and consent boundary unclear, so review is needed to resolve reachability and behavior.
Decision evidence
public snapshot- package.json points to dist/index.cjs and declares no install lifecycle hook.
- dist/mcp.cjs contains helpers that resolve agent skill directories, fetch skill files using a bearer token, and write the files there; their activation and consent boundary remain unresolved.
- The skill file writer rejects absolute paths and paths that resolve outside the selected skill directory.
- The manifest lists only ssh2 and node-pty as runtime dependencies and has no lifecycle scripts.
- The inspected MCP server dispatches tools through a handler, but no caller of the skill-install helper was found in the source search.
Source & flagged code
8 flagged · loading sourceA single source file combines environment access, network access, and code or shell execution; review context before blocking.
dist/mcp.cjsView on unpkg · L4Source contains an obfuscated payload loader that reconstructs and executes hidden code.
dist/mcp.cjsView on unpkg · L1Findings
Affected versions and remediation
This report applies to @whalent/agent-core@0.3.504.
See all versions of @whalent/agent-core for other recorded verdicts.
- Review the evidence and your use of @whalent/agent-core@0.3.504 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.