KT AI Coding Agent Workflows

Reusable workflow templates for AI coding agents.

中文说明见 README.zh-CN.md。

Scope

This repository collects practical workflows for AI-assisted software engineering: planning, code review, debugging, research, release preparation, documentation, and multi-agent task decomposition.

Layout

workflows/
  planning/
  review/
  debugging/
  release/
  multi-agent/

Guidelines

• Keep workflows executable and specific.
• Include inputs, steps, expected outputs, and verification criteria.
• Prefer small templates that can be reused by skills, MCP servers, and CLIs.

Current Templates

Template	Use
planning/repo-optimization.md	Single-repository optimization loop: inventory, targeted reading, edits, validation, commit, and ledger update.
review/pre-landing.md	Pre-landing review focused on correctness, security, regressions, and missing validation.
multi-agent/research-synthesis.md	Multi-agent research synthesis with evidence merging, conflict handling, and decision output.

Verification

git diff --check
rg -n "gho_|Bearer|SERVICE_ROLE|password|cookie|secret|token|API_KEY|/Users/" README.md README.zh-CN.md workflows || true

Sensitive scans may only match safety guidance or placeholders, not real credentials, private paths, or account details.