# Startup Agentic Dev Workflow

> Instructions

- Skill: `uygnoey/startup-agentic-dev-workflow` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add uygnoey/startup-agentic-dev-workflow`
- Raw SKILL.md: https://api.skillmd.com/api/skills/uygnoey/startup-agentic-dev-workflow/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: uygnoey (https://skillmd.com/u/uygnoey)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/uygnoey/startup-agentic-dev-workflow

---


## Instructions
Use this workflow to ship reliably with Claude Code while keeping context clean.

1) Research session
- Use a dedicated session for background research on the feature.
- When needed, use subagents to investigate different parts of the codebase in parallel.
- Produce a single research document that captures what matters.

2) Planning session
- Create an implementation plan with discrete phases.
- Iterate on the plan until it’s coherent and testable.
- Write the plan into a Markdown file you can follow during implementation.

3) Implementation session
- Execute phases systematically.
- If the agent starts producing contradictory or low-quality output, stop early and re-start with a distilled context summary.

4) Context management
- Prefer new sessions for major phases.
- Pass only distilled conclusions forward.

## Examples
- Template for the three-phase workflow: `templates/research-plan-implement.md`
- Template for context distillation between sessions: `templates/context-distillation.md`
- Best practices summary: `references/session-best-practices.md`