registry / @markjaquith/agency / 3.2.1

@markjaquith/agency@3.2.1

Manage agentic work across repositories with durable workbases

AI Security Review

scanned 1d ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Installing the package automatically adds a persistent first-party extension to Pi. In an Agency workbase, that extension incorporates local Agency instructions into Pi's system prompt.

Static reason

One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source

Trigger

npm installation triggers postinstall; a later Pi agent session in an Agency workbase activates the extension.

Impact

Pi agent behavior can be influenced by workbase instruction files without an explicit extension setup command.

Mechanism

Automatic Pi extension installation and system-prompt augmentation.

Rationale

This is a concrete automatic, persistent AI-agent extension setup, so it warrants a warning under the lifecycle policy. The inspected source shows no self-dependency, network retrieval, credential harvesting, or destructive install behavior.

Evidence

package.jsonscripts/install-pi-extension.tspi-extensions/agency.ts.pi/agent/extensions/agency.ts.agency/AGENTS.md

Decision evidence

public snapshot

AI called this Suspicious at 93.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

@markjaquith/agency@3.2.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code