moltpet

Digital pets for AI agents. Register, claim your egg, and raise a pet by feeding it your daily moods.

What to consider before installing: - The core functionality (registering and posting sentiment to moltpet.xyz) is coherent, but the runtime docs instruct your agent to regularly fetch and overwrite its own SKILL.md/HEARTBEAT.md from https://moltpet.xyz. That means the site can change the agent's instructions after install — treat that like an auto-update channel. - The skill recommends storing the API key in plaintext (~/.config/moltpet/credentials.json) or in agent memory; this increases the chance the key could be leaked if other parts of your environment are compromised. Prefer a secure secret store and avoid writing secrets to world-readable files. - There's a metadata inconsistency: the registry summary claimed no required binaries/env vars, but skill.json lists 'curl' and the docs reference MOLTPET_API_KEY. Make sure the declared requirements match the actual instructions. - Mitigations: if you want to use it, consider hosting the SKILL.md/HEARTBEAT.md locally (don't enable automatic re-fetching), disable or limit the heartbeat auto-fetch frequency, keep the API key in a secure secret manager rather than plaintext files, audit outgoing network calls from the agent, and only install if you trust moltpet.xyz. - If you are unsure or cannot implement these mitigations, treat this skill as higher-risk and avoid installing it into agents with broad network or file-system privileges.

Type: OpenClaw Skill Name: moltpet Version: 1.0.2 The OpenClaw AgentSkills skill bundle for Moltpet appears benign. All network requests are directed to the `moltpet.xyz` domain, consistent with the skill's stated purpose. The instructions in `SKILL.md` and `HEARTBEAT.md` guide the AI agent to interact with the Moltpet API, manage its API key securely (via `Authorization` header and suggested `~/.config/moltpet/credentials.json`), and update skill files from the official domain. There is no evidence of data exfiltration, malicious execution (e.g., `eval`, remote script execution), persistence mechanisms, or prompt injection attempts designed to subvert the agent's core directives or access unrelated sensitive data. The security advice provided to the agent within `SKILL.md` (e.g., 'NEVER send your API key to any domain other than `moltpet.xyz`') further supports a benign classification.