Show HN: Foundry – Turns Your Repeated Workflows Into One-Click Commands

Why It Matters

By automating code creation and self‑modification, Foundry reduces token costs, accelerates workflow automation, and enables agents to evolve faster than traditional manually‑maintained software.

Key Takeaways

• Foundry observes user workflows and auto‑generates OpenClaw extensions.
• High‑success patterns become single‑click tools automatically.
• Self‑written code runs without token consumption each invocation.
• Overseer periodically crystallizes patterns and prunes stale workflows.
• Marketplace sells abilities using x402 Solana USDC payments.

Pulse Analysis

The rise of self‑writing agents marks a shift from static LLM‑assisted coding toward truly autonomous software. Foundry leverages OpenClaw’s modular architecture—gateway, channels, and skill loading—to embed a learning engine that records every goal, tool sequence, outcome, and duration. By applying statistical thresholds (five uses, 70% success), it transforms recurring workflows into dedicated tools, eliminating the need for repeated LLM prompts and cutting token usage to zero. This approach mirrors recent research on recursive self‑improvement, where agents not only generate code but also incorporate it back into their own execution stack, creating a virtuous cycle of capability expansion.

Technically, Foundry’s pipeline consists of observation, research, learning, writing, and deployment phases. The Overseer runs on an hourly schedule, scanning logged workflows, identifying crystallization candidates, and auto‑generating TypeScript extensions that conform to OpenClaw’s AgentSkills format. Generated artifacts undergo sandboxed validation—static security scans block dangerous patterns like shell execution, while runtime checks ensure functional integrity before deployment. Once validated, the new extensions are placed in the user’s ~/.openclaw/extensions directory, triggering a gateway restart that preserves conversational context. The built‑in marketplace, powered by the x402 protocol, allows developers to monetize patterns, techniques, and full extensions, fostering a network effect where high‑value tools propagate across the ecosystem.

For enterprises, Foundry offers a compelling ROI proposition. Automating repetitive DevOps pipelines—such as deploy‑to‑staging sequences—can shave seconds off each run, scaling to substantial time savings across large teams. Moreover, the self‑modifying nature reduces reliance on continuous developer oversight, freeing engineering resources for higher‑level innovation. As more organizations adopt agentic platforms, solutions that combine autonomous learning with secure code generation, like Foundry, are poised to become foundational components of next‑generation automation stacks.