registry / @focuslabs/leyline / 0.1.2

@focuslabs/leyline@0.1.2

Meta-harness for Kiro — agent orchestration, persistent memory, task coordination, and self-learning capabilities

AI Security Review

scanned 1h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. An explicit `leyline init` writes first-party Kiro agent-extension configuration in the current project. The optional OpenAI embedding provider POSTs caller-supplied text to the configured embeddings endpoint.

Static reason

No blocking static signals were detected.

Trigger

User runs `leyline init`; an MCP client later starts the configured server or selects the OpenAI embedding provider.

Impact

Adds agent roles including shell-capable templates and registers a project MCP server; optional embeddings transmit supplied memory text to OpenAI.

Mechanism

explicit Kiro extension setup plus optional embedding API requests

Rationale

No source evidence establishes malicious installation-time or covert behavior. The explicit Kiro extension setup and shell-capable agent templates warrant a warning under the agent lifecycle policy.

Evidence

package.jsonbin/leyline.jsdist/commands/init.jsdist/commands/mcp.jsdist/templates/mcp-config.jsdist/templates/agents.jsdist/services/vector/providers/openai.js.kiro/settings/mcp.json.kiro/agents/leyline-orchestrator.json.kiro/agents/leyline-coder.json.kiro/agents/leyline-tester.json.kiro/agents/leyline-reviewer.json.kiro/agents/prompts/*.md.kiro/skills/*/SKILL.md

@focuslabs/leyline@0.1.2

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code