registry / @markjaquith/agency / 2.74.0

@markjaquith/agency@2.74.0

Manage agentic work across repositories with durable workbases

AI Security Review

scanned 14d ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. On npm postinstall, the package copies its Pi extension to ~/.pi/agent/extensions/agency.ts, a globally loaded Pi location. The extension is guarded to Agency workbases but can augment Pi agent prompts there.

Static reason

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

Trigger

npm postinstall followed by Pi activity in a directory with agency.json

Impact

Changes an AI-agent control surface without an explicit setup command; no exfiltration or remote payload chain was found.

Mechanism

global Pi extension installation and guarded prompt augmentation

Rationale

This is a real install-time AI-agent extension mutation, but it is a package-owned, guarded integration rather than a concrete malicious chain. It warrants a warning under the lifecycle control-surface policy, not a publish block.

Evidence

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

Decision evidence

public snapshot

AI called this Suspicious at 92.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

postinstall invokes the Pi extension installer.

@markjaquith/agency@2.74.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code