registry / @markjaquith/agency / 3.1.1

@markjaquith/agency@3.1.1

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 writes a persistent Pi agent extension outside the package. On later Pi agent starts in a validated Agency workbase, that extension reads workbase instructions and modifies the agent system prompt.

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 start in an Agency workbase.

Impact

It changes an AI-agent control surface without an explicit setup command; source inspection found no exfiltration or remote payload execution.

Mechanism

Automatic global agent-extension installation and startup prompt augmentation.

Rationale

This is an automatic first-party agent-extension setup that modifies a global AI-agent control surface, so it warrants a warning. The inspected implementation does not establish malicious behavior or the broader mutation chain required for blocking.

Evidence

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

Decision evidence

public snapshot

AI called this Suspicious at 95.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

@markjaquith/agency@3.1.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code