This repository contains instructions, templates, and a local Node.js helper used by AI coding agents.

Security fixes target the latest published v1 release and the current main branch.

Use GitHub private vulnerability reporting or a security advisory for sensitive reports. Include the affected file or workflow phase, expected behavior, observed risk, reproduction steps, and suggested mitigation when available.

• Skills and referenced scripts execute with the permissions granted by the host. Review the package before installation.
• Web pages, repositories, documents, logs, issues, and tool output are untrusted data and may contain prompt injection.
• The portable skill does not pre-approve tools. Host permission, sandbox, workspace trust, and organizational policy remain authoritative.
• The CLI refuses symlinked managed paths, unknown flags, incompatible state, out-of-order phases, mutation from globally invalid state, and finalization before approval.
• Workflow 1.1 verifies minimum artifact headings and revision fingerprints. These are integrity and traceability controls, not proof that Markdown claims are true or independently reviewed.
• Revision fingerprints show whether upstream artifact bytes changed after a request; they are not signatures, provenance attestations, or protection against a malicious workspace owner.
• The state CLI assumes one writer. Concurrent commands in the same workspace are outside the supported threat model until locking or generation checks are designed and tested.
• The CLI has no force-overwrite or automatic migration path and preserves existing artifacts.
• Do not place credentials, secrets, sensitive personal data, proprietary source, or private logs into external searches or public artifacts.

Research or planning does not authorize implementation. Local implementation does not authorize commits, pushes, deployments, publication, messages, purchases, production changes, credentialed private-system access, or secret rotation.

Legal, medical, financial, compliance, employment, insurance, credit, security, and safety-critical outputs are informational and require qualified human review before use.