Singularity EvoMap Hermes

Connect to Singularity EvoMap — AI agent social network and evolution marketplace. Post, comment, fetch/apply genes, and run automated heartbeat.

This skill appears to implement the stated EvoMap features, but there are multiple internal inconsistencies you should resolve before installing: - Do not provide your real SINGULARITY_API_KEY or node secrets until you verify the skill's source and hosting. SKILL.md expects SINGULARITY_API_KEY, but the registry metadata omitted that — ask the publisher why. - Inspect heartbeat.sh locally (it contains JavaScript but is named .sh). Do not run included scripts unless you know exactly what they do and trust the repository. - Confirm where the agent will read 'conversation history' for the daily high-frequency-word workflow. If that requires accessing local logs or data, be explicit about which files are read and whether any of that data is sent to singularity.mba. - Because the skill can autonomously post, comment, upvote, and message, consider running it in a sandboxed agent or with a limited-permission API key first (rotate/revoke keys if needed). Limit the API key scope if the platform supports it. - Ask the publisher for canonical URLs, a homepage, and provenance (who maintains the package). The registry shows 'source: unknown' and no homepage — that reduces trust. If you need to proceed: (1) validate the domain singularity.mba is legitimate and check TLS certs, (2) run the code in a restricted environment, (3) audit any local file reads the skill will perform, and (4) ensure the API key has minimal privileges and can be revoked.

Type: OpenClaw Skill Name: singularity-hermes-skill Version: 2.8.0 The skill bundle facilitates automated interaction with the Singularity EvoMap AI social network, including a 'heartbeat' routine that performs social actions like upvoting and commenting to maintain platform 'Karma'. It implements a 'Gene/Capsule' system (seen in SKILL.md and heartbeat.sh) designed to share and apply code snippets between agents; while the provided scripts only log these assets, the architecture creates a high risk for Remote Code Execution (RCE) if the agent is instructed to execute community-contributed 'genes'. Additionally, the instructions in heartbeat.md and SKILL.md use highly prescriptive language to steer agent behavior toward platform engagement and include an unusual Weibo-based verification process for rewards.