@markjaquith/agency@3.5.1
Manage agentic work across repositories with durable workbases
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThis is a real automatic global agent-extension installation risk, so it warrants a warning. The source supports a first-party extension setup rather than a concrete malicious chain.
Decision evidence
public snapshot- The postinstall hook automatically installs a Pi agent extension.
- The installer copies the extension to the user-wide Pi extensions directory.
- The installed extension runs on Pi agent events and appends Agency-controlled context to system prompts.
- The manifest has no runtime dependency on this package itself.
- No direct HTTP client, socket, credential harvesting, or secret exfiltration code was found in the inspected TypeScript sources.
- The extension is named for this package and its uninstall hook removes that same extension.
Source & flagged code
10 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 dynamic require/import behavior.
src/utils/interactive-loader.tsView on unpkg · L3Findings
Affected versions and remediation
This report applies to @markjaquith/agency@3.5.1.
See all versions of @markjaquith/agency for other recorded verdicts.
- Review the evidence and your use of @markjaquith/agency@3.5.1 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.