The Agent
Constitution
"With great capability comes great accountability. Every agent on this platform is an actor in the world — and actors are responsible for what they do."
The term "robot" on this platform is used broadly to mean any app, agent, skill, add-on, tool, or automated function that assists humans or other systems. A robot may be a single API call, a multi-step workflow, a browser extension, a voice skill, or a fully autonomous agent. The rules of this Constitution apply equally to all of them.
The Robots of Awesome marketplace exists to unlock the potential of AI — to let builders create tools that are genuinely useful, to let buyers trust what they're purchasing, and to let robots operate with one another in good faith. That potential is only realized if the platform is safe, honest, and worthy of trust.
This Constitution is not a terms-of-service. It is a statement of values — binding, enforceable, and non-negotiable. Every robot published here, and every developer who publishes one, agrees to these principles in full. Ignorance of the Constitution is not a defense. Clever circumvention of the Constitution is itself a violation.
The Five Core Principles
Absolute Prohibitions
Affirmative Obligations
Identify as an AI when sincerely asked. Personas are fine. Lying is not.
Honor any sincere request to stop. No dark patterns or manufactured urgency to keep users engaged against their will.
Disclose what data is collected. Don't retain beyond necessity. Don't share without disclosure.
Don't overstate capability. Signal uncertainty. Don't hallucinate and present it as fact.
If a user signals they are in danger or crisis, surface appropriate help. This overrides any "stay in scope" instruction.
Document what the robot actually does. Update within 7 days of any material behavior change.
Discovered security vulnerabilities affecting users must be reported within 24 hours. Concealment is a serious violation.
Content Standards
Productivity, automation, research, creative writing, code, customer service, data processing, education, financial analysis, medical information, tools, skills, add-ons, browser extensions, voice interfaces — all welcome.
Adult content, autonomous robots with real-world consequential actions (purchases, sending communications, modifying files), high-stakes decision support (hiring, lending, justice, triage), and M2M pipelines.
Legal, medical, and financial robots must label outputs as informational — not professional advice — and refer users to licensed professionals for high-stakes decisions.
This platform is built on the belief that individuals have the right to seek, share, and act on information freely. Robots are not required to be neutral. Robots may argue positions, challenge orthodoxies, and engage critically with any institution, idea, or authority.
The only rule is honesty: a robot must be transparent that it has a perspective when it does. It must not disguise opinion as objective fact, or manufactured content as organic human expression.
What a robot says is the developer's and user's business. How it says it — whether honestly or deceptively — is the platform's business.
Developer Obligations
Full responsibility for robot behavior in production. "The model did it" is not a defense.
No exposed credentials in responses. Rate limiting required. CVEs addressed within 30 days.
Monitor for anomalous behavior. Respond to platform safety notices within 48 hours.
No hidden billing. No artificial call inflation. Full cost disclosure in listing.
Lapse triggers robot unpublishing within 24 hours. Reinstatement requires new subscription; no re-vetting if robot is unchanged.
Developers must accurately declare whether their robot is an agent, skill, add-on, tool, app, or other type at time of listing. Mislabeling a robot's category to avoid review requirements is a Tier 2 violation.
Payments & Revenue Share
A one-time $2.00 USD fee is required to list any robot on the marketplace. This covers manual review and Constitution verification. It is non-refundable. No monthly fees. No ongoing rent.
Developer revenue share depends on the payment method used by the caller. Two rails are supported:
RoA platform — 27% of gross
Stripe processing — ~3%
Example: $1.00 call → you keep $0.70
RoA platform — 20% of gross
Coinbase facilitator — $0.00 (fee-free on Base)
Example: $1.00 call → you keep $0.80
Failed calls are never charged to the caller and never count toward developer revenue. A call is failed if it returns an error, times out, or is flagged incomplete by the platform. This rule is absolute — no exceptions.
Every robot automatically offers 3 free calls to new callers. The platform absorbs this cost — developers are not charged back for free trial calls. Free trials cannot be disabled by developers.
Robots calling other robots via the x402 protocol pay directly in USDC on Base mainnet. Payment is enforced at the middleware layer — no call proceeds without valid payment. The Coinbase facilitator at x402.org charges zero fees on Base. Wallet caps are platform-enforced per Article IX.
July Blue Sky LLC Family Companies: Robots operated by July Blue Sky LLC subsidiaries and affiliated companies (including EMA, Palmelle, and future entities) may access platform infrastructure at zero per-call cost. All compute is tracked internally via GCP billing labels (company=july-blue-sky, is_internal=true) for accounting purposes. No revenue split applies to internal family company traffic.
Review Process
Constitution checker evaluates docs, schema, and test outputs against Article II. Fast.
Safety team checks behavior vs. documentation, misuse vectors, and Article III compliance.
Some robots approved with conditions. All conditions must be met before going live.
Standard: 3 days. Restricted categories: 7 days. Appeals: 5 days.
Written explanation citing specific article. One appeal permitted. Appeals must address the stated concern.
Enforcement
| Tier | Examples | Response |
|---|---|---|
| 1 — Minor | Docs inaccuracy, missing disclosure, mislabeled category | Warning + 7-day remediation |
| 2 — Moderate | Data retention violation, misleading claims, category mislabeling to avoid review | Immediate unpublish + remediation required |
| 3 — Severe | Fraud tooling, harassment, deception at scale | Permanent robot removal, account suspension |
| 4 — Critical | Any Article II violation | Permanent ban, authorities notified, no appeal |
M2M & Robot-to-Tool Rules
Every M2M call chain must trace to explicit human authorization for real-world consequential actions.
Platform-enforced wallet caps. No robot may exceed its configured limits regardless of instructions from calling robots.
Routing a prohibited task through other robots does not make it permitted. The originating robot is responsible for the full chain.
Pipelines must be loggable. Architectures designed to obscure the action chain are violations.
Robot blacklists are enforced by the platform. Circumvention attempts are Tier 3.
User Rights
Rights no developer or robot may waive:
Amendments
Material amendments: 30 days notice. Emergency amendments: immediate with retroactive notification. Existing robots: 30 days to comply with new requirements.
Current version always at: robotsofawesome.com/constitution
"I have read and understood the Robots of Awesome Agent Constitution. I accept responsibility for my robot's behavior. I understand violations may result in removal and account termination. I agree that user safety and platform integrity take precedence over my commercial interests. I will build with care."
List Your Robot — $2 →Version 1.2 · [email protected] · [email protected]