Nostr Boutique Update #2
- The Stack
- Agent 1: Development & Deployment
- Agent 2: Promo Agent
- Promo Agent Responsibilities
- The Shared Loop
- Why This Matters
- The Vision for Nostr.boutique
- Final Thought
I’ve been experimenting with a new workflow for building and growing Nostr.boutique using an agentic development & promo setup.
The idea is simple: instead of treating development, deployment, promotion, feedback, and community engagement as separate disconnected tasks, I want to connect them into one continuous loop.
For this setup, I’m thinking about two agents working together:
- A Development & Deployment Agent
- A Promo Agent connected to a wallet using NWC

The goal is to create a lightweight cycle where features can move from idea → build → deployment → outreach → feedback → back into development.
The Stack
This setup is designed around Nostr-native tooling:
- Nsite’s for publishing and hosting
- Ngit for git-based development workflows
- NWC for giving the promo agent wallet access in a controlled way
- Nostr as the social, discovery, and feedback layer
The bigger idea is to make Nostr.boutique feel like a living client project that can continuously ship, listen, promote, and improve.
Agent 1: Development & Deployment
The first agent is responsible for the development side of the cycle.
This agent takes in issues, feature requests, and ideas, then turns them into an actionable development flow.
1. Take Issues & Features
The cycle starts with input.
This can include:
- Bugs
- Feature requests
- UX improvements
- Community feedback
- Ideas from users
- Issues reported by the promo agent
The agent collects these inputs and treats them as potential development tasks.
2. Turn Them Into a Code Action Plan
Once the issues or features are collected, the agent turns them into a clear code action plan.
This is where the task becomes more structured:
- What needs to be built?
- What files or components might be affected?
- What is the expected outcome?
- What should be tested?
- What risks or tradeoffs exist?
This step should include a human guidance window.
That means the agent does not just run off and build blindly. There is a moment where a human can review, steer, correct, or refine the plan before code is written.
3. Create a Branch for the Build
After the action plan is reviewed, the agent creates a dedicated branch for the build.
This keeps the implementation isolated and easier to inspect.
This step should also include a human guidance window, giving humans another chance to review direction before execution.
The branch becomes the workspace for the feature, fix, or experiment.
4. Deploy & Trigger the Promo Agent
Once the build is complete and deployed, the development agent triggers the second agent: the promo agent.
This is where the cycle shifts from building to outreach.
The deployment is not the end of the process. It becomes the starting signal for promotion, engagement, and feedback gathering.
Agent 2: Promo Agent
Currently the agent doing the work is: @f7fg…35sy
The second agent is focused on outreach, community interaction, engagement monitoring, and reporting feedback back into development.
This agent would be set up with a wallet using NWC, allowing it to participate in Nostr-native promotion and engagement flows.
Wallet Setup with NWC
The promo agent should have controlled wallet access through Nostr Wallet Connect.
The goal is not to give an agent unlimited financial freedom, but to give it a bounded ability to interact with the Nostr ecosystem.
For example, it could use small zaps for lightweight promotional activity, engagement experiments, or community interaction.
Promo Agent Responsibilities
Outreach & Zapvertisement
The promo agent can run continuous outreach.
One idea I’m exploring is random or semi-random zapvertisement: small Nostr-native promotional actions intended to keep visibility and engagement alive.
This could be used to:
- Highlight new releases
- Bring attention to useful features
- Support community discovery
- Encourage people to try Nostr.boutique
- Keep the project visible without requiring manual posting every time
Engagement Monitoring
The promo agent can also monitor simple engagement signals.
For example:
- Are people reacting to releases?
- Are users replying with bugs or questions?
- Are certain posts getting more traction?
- Are specific features creating more discussion?
- Are there repeated requests or complaints?
This does not need to be overly complex at first. Even basic monitoring can create useful feedback.
Customer Support
The promo agent can help with lightweight customer support.
This could include answering common questions, pointing users to documentation, or collecting support issues that need human or developer attention.
The agent does not need to pretend to know everything. A good support agent should know when to escalate.
Report Issues
One of the most important roles of the promo agent is to feed issues back into the development cycle.
If users report problems, confusion, missing features, or repeated friction, the promo agent can turn that into structured input for the development agent.
This creates the loop:
users → promo agent → development agent → deployment → users
Announce Releases & Monthly Stats
The promo agent can also announce new releases and publish regular updates.
This could include:
- New feature announcements
- Bug fix summaries
- Monthly usage stats
- Engagement summaries
- Most requested features
- Development progress updates
This helps make the project feel active, transparent, and alive.
The Shared Loop
The most important part of this setup is the intersection between the two agents.
The development agent and promo agent are not separate silos. They overlap through a shared feedback loop.
That shared loop looks like this:
- The development agent ships a release.
- The promo agent announces and promotes it.
- Users engage, react, ask questions, or report issues.
- The promo agent monitors and collects that feedback.
- The feedback becomes new issues or features.
- The development agent turns those into the next build.
In simple terms:
Continuous build → launch → outreach → feedback → build again
This is the cycle I want Nostr.boutique to move toward.
Why This Matters
Nostr projects often need more than just code.
They need:
- Continuous shipping
- Better visibility
- Feedback loops
- Community support
- Release communication
- Sustainable promotion
Agentic workflows can help connect these pieces together.
Instead of only using agents to write code, I’m interested in using agents as part of the full project lifecycle.
That means agents can help with:
- Planning
- Development
- Deployment
- Promotion
- Support
- Feedback collection
- Reporting
For a Nostr-native project, this feels especially interesting because the same network used for social discovery can also become part of the development feedback loop.
The Vision for Nostr.boutique
With Nostr.boutique, I want to experiment with a more self-sustaining development cycle.
The project should be able to:
- Collect ideas from users
- Turn those ideas into development plans
- Let humans guide the important decisions
- Build in branches
- Deploy through Nsite’s and Ngit
- Trigger Nostr-native outreach
- Use NWC-powered wallet interactions carefully
- Monitor feedback
- Report issues back into development
- Announce progress and monthly stats
This is not about removing humans from the process.
It is about creating better loops where humans can guide, steer, and supervise while agents handle more of the repetitive coordination work.
The human stays in the important decision points.
The agents keep the cycle moving.
Final Thought
The setup I’m exploring for Nostr.boutique is a simple two-agent system:
- Agent 1 builds and deploys
- Agent 2 promotes, supports, monitors, and reports back
Together, they form a continuous development and engagement loop.
For Nostr client development, especially using tools like Nsite’s, Ngit, and NWC, this kind of workflow could make it easier to ship faster, listen better, and keep the community engaged.
That’s the so far the new reporting 🫡
Write a comment