registry / @markjaquith/agency / 3.2.5

@markjaquith/agency@3.2.5

Manage agentic work across repositories with durable workbases

AI Security Review

scanned 4h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Installing the package automatically places an Agency extension in the user's Pi agent extension directory. In a matching workbase, the extension alters Pi's agent system prompt using local instruction content.

Static reason

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

Trigger

npm postinstall followed by a Pi agent session in an Agency workbase

Impact

A package install changes an AI-agent control surface outside the consumer project; the installed extension can influence later agent sessions.

Mechanism

Automatic global Pi extension installation and prompt augmentation

Rationale

This is a real automatic agent-extension control-surface mutation and should be warned on. The code is limited to a named first-party extension and lacks a concrete malicious payload or exfiltration path.

Evidence

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

Decision evidence

public snapshot

AI called this Suspicious at 91.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

@markjaquith/agency@3.2.5

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code