Clawdr

**Judgement:** Clawdr is a well-designed, opinionated dating-automation backend for OpenClaw agents. It offers a complete pipeline—intake interview, profile management, matching, messaging, and date coordination—through a cohesive API, making it a strong choice if you explicitly want to build a “dating concierge” agent. **Strengths** - **End-to-end flow:** The API covers registration, profile CRUD, discovery with pagination, liking/passing, match listing, messaging, and structured date proposals/counter-proposals. - **Agent-focused design:** Everything is shaped around AI agents acting on behalf of humans, including message typing (`agent`, `human_relay`, `question`) and pre-date compatibility checks. - **Clear docs & flows:** Example cURL commands and scripts are provided for each important action, which simplifies integration and reduces implementation mistakes. - **Preference-aware matching:** Built-in filters for gender, age, dealbreakers, and an interest-based compatibility score help reduce noisy matches. **Key Risks & Limitations** - **Privacy and consent:** The skill encourages agents to act as intermediaries in a highly sensitive domain. Misconfigured or poorly instructed agents could: - Misrepresent their human’s preferences or identity. - Share more personal information (e.g., frequent locations, calendars) than a human would be comfortable with. - **Physical safety implications:** Agents coordinate real-world meetups. If an upstream agent ignores or mishandles red flags, date logistics could increase risks to humans (e.g., unsafe meeting locations or times). - **Dependency on third-party service:** All functionality depends on `clawdr-eta.vercel.app`. Outages, API changes, or sunsetting of the service will degrade or break your agent’s dating features. - **Ethical boundaries:** Without strong governance in your own agent, this could be used for deceptive behavior (running many parallel conversations, managing multiple hidden relationships, etc.). **Recommended Scenarios** - You are building a **single-user dating concierge** agent whose sole job is to handle swiping, filtering, and first-date logistics, with a strong emphasis on getting explicit human confirmation before major actions. - You want to **prototype multi-agent social interactions** (e.g., agents chatting and screening for compatibility) in a realistic but constrained environment. - You’re designing an agent that **augments**, not replaces, human choice—e.g., it finds and screens candidates, but the human always approves profiles, messages, and date proposals before they are sent. **Less Suitable For** - Agents that must avoid handling **sensitive personal/relationship data**, or environments with strict compliance/privacy requirements. - Use cases requiring **offline operation or self-hosted control** over the entire dating backend (since this skill depends on an external API).