registry / @cursor/july / 0.2.22

@cursor/july@0.2.22

Build, test, and deploy BDK bots and agents as code on Cursor's agent harness.

AI Security Review

scanned 11h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Package installation automatically copies package-owned skills into the user's Cursor skills directory. Slack continuation handling can also write a carrier script and register Cursor hooks in the agent workspace.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

npm install triggers the skill copy; a Slack continuation with a continuation key and disabled steering can trigger workspace hook setup.

Impact

These behaviors extend package-owned agent functionality into Cursor's skill and hook mechanisms. The inspected code did not establish credential theft, remote code execution, or destructive effects.

Mechanism

The installer copies BDK skills into its namespace. The Slack carrier writes supporting files and adds pre-tool and post-tool hook entries that invoke the carrier.

AI rationale

The source shows an automatic first-party skills installation and guarded Slack workspace hook setup. These are agent integration lifecycle risks, but the inspected behavior does not establish a concrete attack, so the package warrants a warning rather than a block.

Evidence

package.jsondist/internal/install-cursor-skills.jsdist/channels/slack/nudge-carrier.jsdist/internal/session-engine.js

Decision evidence

public snapshot

AI called this Suspicious at 84.0% confidence as Dangerous Capability with false-positive risk.

@cursor/july@0.2.22

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code