registry / @sema-agent/core / 1.317.0

@sema-agent/core@1.317.0

Stateless, task-oriented AI agent core

AI Security Review

scanned 17h ago · by lpm-firewall-ai

No confirmed malicious attack surface. Runtime shell, web, workflow, and memory features require explicit consumer configuration or invocation; no install-time mutation or package-initiated exfiltration was found.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Explicit consumer use of execution, web, workflow, or memory APIs

Impact

No concrete unconsented payload execution, credential exfiltration, or foreign agent-control mutation established

Mechanism

Opt-in AI-agent utilities with guarded web fetching and package-owned memory persistence

Rationale

The scanner's reverse-shell and eval signals resolve to security-prompt/denylist text and explicit workflow capabilities, not package-initiated attack behavior. Source inspection found no consumer-install lifecycle execution, covert endpoint, exfiltration path, or foreign AI-agent control-surface mutation.

Evidence

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

Decision evidence

public snapshot

@sema-agent/core@1.317.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code