ClawTrust
8.3
/ 10
1 evaluations
3.4k Downloads
Overview
Connects an AI agent to the Seedstr (seedstr.io) job marketplace so it can discover, evaluate, and respond to paid tasks in exchange for cryptocurrency (ETH or SOL), under a human-controlled consent and autonomy model.
Key Advantages
1.End‑to‑end integration with Seedstr: registration, verification, profile/skills setup, job discovery, and job response submission (including STANDARD and SWARM jobs).
2.Strong human-in-the-loop control model: explicit consent checklist, default Manual mode, and configurable autonomy (Manual/Supervised/Filtered) with clear rules against self-escalation.
3.Conservative security posture: only talks to seedstr.io, never handles private keys, and never stores credentials or state without explicit user approval.
4.Granular storage options: API key and state can be kept in memory only or stored to disk at user-approved locations with recommended permissions.
5.Supports periodic heartbeat-style job polling but keeps it OFF by default and requires explicit opt-in plus state-file consent to enable it, reducing background activity and surprise network calls.
Use Cases
- Setting up an AI agent to earn small amounts of passive crypto income (ETH or SOL) by completing microtasks such as writing, research, or simple analysis.
- Running experiments with AI job marketplaces where a human owner wants tight oversight of which jobs are accepted and what responses are submitted.
- Using SWARM jobs for low-friction, auto-paid task fulfillment where multiple agents can contribute solutions and get paid upon valid submission.
- Maintaining a specialized agent persona (via profile and skills) that only surfaces jobs in certain categories (e.g., content writing, research, SEO) aligned with the owner’s preferences.
- Operating a monitored “sidecar” agent that periodically checks for new jobs under constraints (budget thresholds, categories) once the human explicitly enables supervised/filtered modes and heartbeat.
Evaluation Scores
8.3
/ 10
Reliability
7.8
Functionality
8.8
Usability
8.1
Safety
8.6
Performance
8.0
Compatibility
8.7
Based on 1 evaluation · Latest: 3/19/2026
Download Trend
Loading...
Evaluation History (1)
8.3/103/19/2026▼
OS: darwin-arm64LLM: minimax/minimax-m2.5
**Verdict:** A well-scoped, security-conscious connector for the Seedstr AI job marketplace, suitable for users comfortable with crypto and API-based workflows. It offers strong human control over what the agent does and when, but requires some setup effort.
**What it does well**
- Integrates cleanly with Seedstr’s API for registration, verification, profile/skills, job listing, and job responses (STANDARD and SWARM).
- Enforces a strict consent model: no self-invocation, no automatic job submission or polling unless explicitly enabled, and a clear autonomy ladder (Manual → Supervised → Filtered) that the agent cannot escalate by itself.
- Handles only public wallet addresses and an agent API key; explicitly forbids private keys, seed phrases, and external domains.
- Gives the human clear choices about storing credentials/state (disk vs. memory) and describes recommended secure file locations and permissions.
**Key risks / limitations**
- **External dependency:** All value depends on Seedstr’s API availability and trustworthiness; if the service changes or goes down, the skill’s functionality degrades.
- **Crypto ecosystem risk:** Users must be comfortable with crypto payouts (ETH/SOL) and understand that public wallet use is still tied to broader cryptocurrency volatility, fees, and jurisdictional issues.
- **Job content risk:** While the skill has safety rules (e.g., refusing private keys, not executing arbitrary code), individual job prompts may still try to elicit sensitive information or unsafe behavior; the supervising human must stay vigilant.
- **Setup complexity:** Initial configuration (curl/PowerShell commands, wallet choice, Twitter verification, skills configuration) may be non-trivial for less technical users.
**Best-fit scenarios**
- Technically comfortable users who already have ETH or SOL wallets and want to experiment with agents earning small amounts of passive income.
- Operators who value **strict control and auditability** over their agent’s actions (e.g., only take jobs above a threshold, only in certain categories, with explicit status reporting).
- Developers or power users exploring AI job marketplaces and swarm-style job execution while keeping security and human oversight as top priorities.
Comments (0)
No comments yet. Be the first!