The "Claude Code" PRD Workflow: A Secret Weapon for Technical PMs

TL;DR

• Hybrid Efficiency: Use targeted API scripts for "bulk" data transport (precision + token savings) and Claude Code/MCP for reasoning and refinement.
• Offline-First: Break free from slow, distracting web editors. This workflow allows you to stay focused in your preferred IDE (Cursor, VS Code) or AI environment (Claude Code, Antigravity) while drafting in Markdown.
• Safety First: Always use sandbox environments and "Review Diffs" mindsets to avoid catastrophic bulk overwrites.

---

As a Technical Product Manager, your job sits at the intersection of high-level strategy and deep technical implementation. The tools of the trade—often a mix of Jira, Confluence, and Slack—can feel disconnected from the actual engineering reality. It's time to borrow a page from the developer's handbook: Treat your requirements like code.

Enter the Claude Code PRD Workflow.

The Problem: The Browser & Context Bottleneck

Drafting a PRD usually involves:

1. Opening Confluence and wrestling with its WYSIWYG editor.
2. Juggling 15+ tabs of technical documentation and Jira tickets.
3. Manually summarizing technical constraints from Slack or GitHub.

This process is not only slow but also prevents you from getting the most value out of an LLM. Simply dumping raw, unstructured data into an AI leads to token waste and lower-quality outputs.

The Solution: The "Hybrid" Pull/Push Workflow

The most efficient version of this workflow doesn't rely only on AI agents for data transport. Instead, it uses a Hybrid approach:

1. API Scripts: For precise "Pulling" and "Pushing" of raw data (saves tokens, ensures formatting).
2. Claude Code + MCP: For intelligent reasoning, cross-referencing, and drafting.

Step 1: "Pull" Your Context (Precision over Agents)

While you can ask an AI agent to "Find and fetch the docs," it's often more efficient to use a simple script targeting the Jira REST API.

The Hybrid Move: Use a script to handle complex filtering conditions and precisely control the fetch scope. This is where API scripts shine—mapping specific Jira issue types, recursive child page fetching, or filtering by custom fields that might confuse an agentic search.

# Example: Precision pull with complex flags
python scripts/fetch_jira.py --jql "project = 'AUTH' AND status = 'In Progress' AND labels in ('v2-refactor')" --depth 2 --output target_scope.md

By handling the "transport" via API, you provide Claude with a clean, pre-filtered context. This ensures accuracy when the scope is non-trivial and saves thousands of tokens that an agent might waste trying to "navigate" or "guess" the correct set of files.

Step 2: Intelligent Cross-Referencing & Drafting

This is where the magic happens. You're not just editing text; you're developing requirements with an intelligent partner that has "eyes" across your entire stack.

The Multi-Source Context: Use Claude Code + MCP to cross-reference data pulled from Jira and Confluence.

• Prompt: "Claude, analyze the user-stories I just pulled from Jira and cross-reference them with the architecture-spec on Confluence. Identify any technical gaps in our proposed API auth flow."

Advanced Technique: "Portable RAG" with NotebookLM When your technical documentation grows into the hundreds of pages, even the largest context window can struggle.

• The Strategy: Use NotebookLM as your external Brain. Index massive PDF specs, legacy confluence exports, and Slack transcripts.
• The Bridge: Use tools like the NotebookLM Claude Skill to query your indexed knowledge base directly from your CLI. This allows you to "Chat with your Documentation" and pull precise snippets into your PRD draft without manual searching.

Step 3: "Push" Updates (The Safety Protocol)

⚠️ Safety First: Updating production Confluence or Jira via automated tools carries the risk of bulk overwrites. If your script or agent fails to parse a block correctly, you could accidentally wipe out months of collaborative history. ⚠️ 安全至上：使用自動化工具更新生產環境的 Confluence 或 Jira 具有極高的風險。如果解析出現故障，可能會意外抹除數月來的協作歷史。

Standard Safety Operating Procedure (SOP):

1. Sandbox First: Always run your "Push" scripts against a trial Jira project or an orphaned Confluence space first.
2. Review Diffs: Treat documentation updates like Code Reviews. Use a tool to see exactly what changed between your local requirements_v2.md and the cloud version before committing the push.
3. Draft Mode: When possible, push as a "Draft" or "Comment" rather than overwriting the main page content.

Common Pitfalls & How to Avoid Them

• ADF Formatting: Confluence uses Atlassian Document Format (ADF). Converting Markdown -> ADF can be messy. Use robust libraries like atlassian-python-api to handle the heavy lifting.
• Context Window: For massive PRDs (50+ pages), don't dump it all into one prompt. Use RAG or index specific sections using Anthropic's MCP.
• Permissions: Ensure your API tokens have "Least Privilege"—only grant write access to the specific projects or spaces you are managing.

Authoritative Resources

• Anthropic Academy: Best practices for engineering with Claude.
• Atlassian Rovo MCP Server: The official bridge between Claude and your Atlassian data.
• Jira Cloud API Reference: Documentation for building custom precision scripts.
• Model Context Protocol: The open standard for AI data integration.

By bridging the gap between product requirements and the tools developers actually use, you don't just write better PRDs—you build better products.