@markjaquith/agency@3.1.0
Manage agentic work across repositories with durable workbases
AI Security Review
scanned 14d ago · by lpm-firewall-aiThe postinstall hook performs first-party package-owned agent extension setup, which warrants a warning under the agent integration policy. The inspected code does not establish a broader attack chain or malicious payload.
Decision evidence
public snapshot- The postinstall hook automatically runs the package's Pi extension installer.
- The installer copies the bundled extension into the user's Pi agent extensions directory.
- The extension reads workbase instructions and adds them to Pi's system prompt when an agent starts.
- IntegrationService also targets workbase instruction and OpenCode configuration files; its agent setup is part of the package's integration behavior.
- The extension obtains context through the package's own agency command and uses that context to locate workbase instructions.
- The inspected extension behavior shows no credential collection, remote endpoint, or remote payload execution.
Source & flagged code
11 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@3.1.0.
See all versions of @markjaquith/agency for other recorded verdicts.
- Review the evidence and your use of @markjaquith/agency@3.1.0 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.