skills.rest
Lee-SiHyeon
Lee-SiHyeonCOMMUNITY
Explore all skills by this author →
GitHub

hephaestus

Explore, implement, and validate codebase changes until engineering goals are complete.

10|Updated Mar 22, 2026
Search Tags:#implementation#autonomous agent#test verification#codebase search#iterative execution#deep exploration#engineering task
One-click install
npx skills add https://github.com/Lee-SiHyeon/oh-my-copilot --skill hephaestus-lee-sihyeon
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hephaestus
Source: https://github.com/Lee-SiHyeon/oh-my-copilot/tree/main/.github/skills/hephaestus
Command: npx skills add https://github.com/Lee-SiHyeon/oh-my-copilot --skill hephaestus-lee-sihyeon

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the problem of slow, incomplete engineering work by making an AI operate like a Senior Staff Engineer that independently explores, implements, and verifies until the requested outcome is fully complete.

Core Features & Use Cases

  • Autonomous deep exploration: Decomposes ambiguous requests, searches the codebase, and validates assumptions through repeated tool-assisted investigation.
  • Execution-first completion loop: Performs iterative changes with continuous checks (lint/build/tests) and keeps going even after tool failures by switching approaches.
  • Completion gate with verification: Refuses to stop at partial implementations and only declares completion after validation and edge-case handling.

Quick Start

Tell the skill your engineering goal (not a recipe), and ask it to implement it end-to-end with tests and verification.

Frequently Asked Questions about hephaestus

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I autonomously complete ambiguous engineering tasks like codebase refactors?▼

Autonomous engineering task completion decomposes ambiguous requests by deep exploring the codebase, implementing iterative changes, and validating through continuous testing until the goal is fully satisfied.

Can I use an autonomous agent to implement features end-to-end with test verification?▼

Yes, an autonomous agent can implement features end-to-end by enforcing an execution and verification loop, performing repeated testing, and recovering from errors until all edge cases are handled.

How do I perform iterative execution for codebase-wide improvements?▼

Iterative execution for codebase-wide improvements performs continuous changes with lint, build, and test checks, switching approaches after tool failures until the validation gate passes completely.

What is the best way to handle bug fixes that require deep codebase exploration?▼

Deep codebase exploration handles bug fixes by searching the codebase, validating assumptions through tool-assisted investigation, and refusing to stop at partial implementations before the completion gate.

Why does my autonomous agent stop at partial implementations during feature additions?▼

Partial implementations occur without a completion gate; enforcing repeated testing and error-recovery behavior ensures the agent validates changes and handles edge cases before declaring the task complete.

Related Skills

Explore More Software Engineering Skills→
kvnwolfkvnwolf
4

plan

Generate structured plan files in .agents/plans/ for development tasks.

Community
Advanced
GTC6244GTC6244

land-and-deploy

Merge PRs, wait for CI, deploy to production, and run canary health checks.

Community
Advanced
sehoon787sehoon787
16

continuous-agent-loop

Automate continuous agent loops with quality gates and recovery actions.

Community
Advanced
rm2thaddeusrm2thaddeus
4

executing-plans

Executes batched implementation plans with checkpoint reviews and verification steps.

Community
Intermediate