Skip to content
View yeliang-wang's full-sized avatar
0 followers · 3 following
  • 22:11 (UTC -12:00)

Achievements

Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: YOLOAchievement: Quickdraw

Block or report yeliang-wang

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
yeliang-wang/README.md

yeliang-wang

I build production-grade AI agent workflows, AgentOps tooling, and self-evolving product systems.

My current focus is turning AI coding from one-off prompt execution into repeatable operating loops: goal-driven agents, real E2E validation, evidence collection, release governance, and human review gates.

Current Work

  • Octopus AgentOps - Reusable Claude Code and Codex subagent workflows with loop-ready contracts, generated distributions, validation, install tooling, and a portable diagnostics sandbox.
  • EvoPilot - Experiments around production lifecycle governance, release evidence, and long-running agent validation.
  • touge-writing-reboot-skill - A reusable writing and reboot skill distilled from public technical-writing patterns.
  • hermes-agent - Agent engineering experiments around workflow execution and productized automation.

Engineering Direction

I care about agent systems that can prove what they did.

That means:

  • goal-driven loops instead of isolated prompts
  • MCP and UI-level E2E validation instead of API-only demos
  • production lifecycle governance with explicit GO / NO-GO decisions
  • evidence bundles, logs, screenshots, and reproducible checks
  • safe SCM / CI / release handoffs
  • self-evolution proposals that stay behind human review gates

Principle

If an agent cannot produce evidence, respect safety gates, and improve through a repeatable loop, it is not production-ready.

Pinned Loading

  1. hermes-agent hermes-agent Public

    Forked from NousResearch/hermes-agent

    The agent that grows with you

    Python

  2. octopus-agentops octopus-agentops Public

    Reusable Claude Code and Codex agent workflows with a portable diagnostics sandbox.

    Python 1

  3. EvoPilot EvoPilot Public

    Evidence-driven SaaS control plane for AI-agent product evolution, approvals, and release decisions.

    JavaScript 1

  4. touge-writing-reboot-skill touge-writing-reboot-skill Public

    Headless writing and reboot skill distilled from Touge Kancode public writing

    Python 1

  5. ProofOps ProofOps Public

    Evidence-first release governance for AI agent products: real-boundary maturity gates, active stability, and product-native GO/NO-GO decisions.

    Python 1