@markjaquith/agency@3.19.0
Manage agentic work across repositories with durable workbases
AI Security Review
scanned 12d ago · by lpm-firewall-aiThe automatic postinstall installs a first-party agent extension, which is a guarded agent-extension lifecycle risk under the review policy. The source shows ownership checks for existing files and no concrete malicious behavior in the inspected paths.
Decision evidence
public snapshot- The postinstall hook automatically runs the extension installer.
- The installer targets the user’s Pi agent extensions directory.
- Installation creates the destination directory and copies the bundled extension when no file exists there.
- Uninstall removes the extension only when its ownership hash matches.
- Integration inspection also targets project-level agent and OpenCode configuration paths.
- The installer skips an existing destination it cannot verify as package-managed.
- No evidence inspected here shows credential collection, exfiltration, or a broader agent-control write.
Source & flagged code
8 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@3.19.0.
See all versions of @markjaquith/agency for other recorded verdicts.
- Review the evidence and your use of @markjaquith/agency@3.19.0 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.