registry / @markjaquith/agency / 2.71.5

@markjaquith/agency@2.71.5

Manage agentic work across repositories with durable workbases

AI Security Review

scanned 15d ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Installation is configured to invoke a Pi-extension installer, but that installer is not present in the package archive. The bundled extension can alter Pi agent context only when Pi loads it in a discovered Agency workbase.

Static reason

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

Trigger

npm postinstall; subsequently, Pi agent startup in an Agency workbase

Impact

Potential agent-extension lifecycle risk; no verified install-time write, network exfiltration, or remote payload execution.

Mechanism

first-party Pi extension hook and prompt/skill-path injection

Rationale

A postinstall intended to install a first-party agent extension warrants a warning under the lifecycle policy, but the archive lacks the installer and source inspection found no concrete exfiltration, destructive action, or remote payload execution.

Evidence

package.jsonpi-extensions/agency.tsscripts/install-pi-extension.ts

Decision evidence

public snapshot

AI called this Suspicious at 78.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

package.json declares an npm postinstall invoking a Pi-extension installer.

@markjaquith/agency@2.71.5

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code