registry / @co-engram/core / 0.2.1

@co-engram/core@0.2.1

Host-agnostic team memory engine with neuroscience-inspired plasticity.

AI Security Review

scanned 14h ago · by lpm-firewall-ai

A user or host invoking engram_doctor on a Git-backed data root can install Git execution hooks and merge-driver configuration. This is explicit package functionality, but it persists execution into future Git merges.

Static reason

No blocking static signals were detected.

Trigger

Invoke the engram_doctor tool with a data root inside a Git repository.

Impact

Writes Git configuration and executable hook code that runs co-engram after future merges.

Mechanism

automatic Git merge-driver and post-merge-hook installation

Rationale

The package is not malicious by the stated boundary because no npm lifecycle hook or unconsented broad AI-agent control-surface mutation exists. Its automatic Git-hook persistence is a concrete guarded lifecycle risk.

Evidence

package.jsondist/storage/infra-doctor.jsdist/merge/auto-onboard.jsdist/merge/onboard.jsdist/merge/post-merge-hook.jsdist/merge/driver-llm.js~/.co-engram/merge-driver.cjs<repoRoot>/.git/config<repoRoot>/.git/hooks/post-merge<repoRoot>/.gitattributes

Decision evidence

public snapshot

AI called this at 95.0% confidence with false-positive risk.

@co-engram/core@0.2.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code