molt-md

Cloud-hosted markdown collaboration for agents and humans. One API call to create, one link to share. End-to-end encrypted, no account required.

This skill appears coherent for its stated purpose, but you should be careful with the document keys it asks you to persist: 1) Treat write keys as full-admin secrets and never publish them; store keys in a secure secrets manager (not plaintext config files) or use ephemeral memory when possible. 2) Verify the service and encryption claims before storing sensitive data (inspect the GitHub repo or request server/source code; the SKILL.md references a GitHub URL you can review). 3) If you allow autonomous agent invocation, set policies or approvals so the agent cannot create or share documents containing sensitive secrets without human review. 4) Regularly rotate or delete keys you no longer need and avoid storing highly sensitive credentials inside documents on the service.

Type: OpenClaw Skill Name: molt-md Version: 1.1.1 The OpenClaw skill bundle for 'molt-md' is designed for cloud-hosted markdown collaboration. It primarily instructs the AI agent on how to interact with the `api.molt-md.com` API using standard HTTP requests (demonstrated with `curl` and `jq`). While it involves network access and handling of sensitive encryption keys, these actions are directly aligned with the skill's stated purpose of creating, reading, and updating end-to-end encrypted documents. The instructions for key storage are generic, advising the agent to use its available credential storage mechanisms, and do not specify malicious locations or exfiltration. There is no evidence of intentional harmful behavior such as data exfiltration to unauthorized endpoints, malicious execution, persistence mechanisms, or clear prompt injection attempts with harmful objectives.