registry / @sema-agent/core / 1.330.0

@sema-agent/core@1.330.0

Stateless, task-oriented AI agent core

AI Security Review

scanned 1h ago · by lpm-firewall-ai

No confirmed malicious attack surface. Runtime agent capabilities are exposed only through explicit API or CLI use, and the inspected web tool includes SSRF and redirect controls.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Explicit application integration, agent tool invocation, or `sema-tb` CLI execution; none at install or import time.

Impact

No unconsented installation-time mutation, persistence, or exfiltration was established.

Mechanism

User-invoked agent orchestration, web fetch, and execution-environment interfaces.

Rationale

The scanner's reverse-shell signal comes from a static security-monitor prompt, while its shell/network signal in `auto-promote` is a denylist. The remaining execution and network primitives are package-aligned, explicitly invoked runtime features with inspected controls and no concrete malicious chain.

Evidence

package.jsondist/core/auto-mode-prompt-assets.jsdist/core/auto-promote.jsdist/tools/web.jsdist/orchestration/dev-vm-script-runner.jsdist/engine/execution-env/node-execution-env.js

Decision evidence

public snapshot

AI called this Clean at 94.0% confidence as Benign with low false-positive risk.

@sema-agent/core@1.330.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code