cross-repo-prd-writer

Cross-Repo PRD Writer

Generate PRD documents for capabilities that need to be built in peer repositories.

Environment

• CLOSEDLOOP_WORKDIR: The working directory (set via systemPromptSuffix)

Process

Step 1: Load Discovery Results

Read $CLOSEDLOOP_WORKDIR/.cross-repo-needs.json for the list of needed capabilities.

Read $CLOSEDLOOP_WORKDIR/.discovery-cache/{peer_name}.json for each peer to see verification results.

Step 2: Filter Out Local Repos

Before evaluating capabilities, inspect each peer entry in .cross-repo-needs.json:

• If the peer has local: true → exclude it entirely from all further processing. These are add-dir repos whose tasks are already incorporated into the plan; there is nothing to build in a remote repo.
• Only process peers where local is absent or local: false.

Track the count of excluded peers for the output summary.

Step 3: Identify Missing Capabilities

For each non-local peer's capabilities, check the discovery cache:

• If exists: true → skip (already exists in peer)
• If exists: false → include in PRD

Step 4: Generate PRDs

For each non-local peer with missing capabilities, create $CLOSEDLOOP_WORKDIR/cross-repo-prd-{peer_name}.md.

Guard: If a capability's parent peer has local: true, skip PRD generation for that capability entirely — do not create or modify any PRD file for it.

IMPORTANT: Before writing any file, you MUST first attempt to Read it. This is required by Claude Code's safety system:

1. Try to Read the target file path
2. If it exists → you've now read it and can Write
3. If it doesn't exist → the Read will fail, but then Write will work for the new file

# Cross-Repo Requirements: {peer_name}

Generated: {timestamp}
Source Repo: {current_repo_name}

## Context

This document describes capabilities needed from **{peer_name}** to support implementation in **{current_repo_name}**.

## Required Capabilities

### 1. {capability_description}

**Type:** {endpoint|model|component|service}
**Needed by tasks:** {T-X.X, T-X.Y}

**Description:**
{Business-level description based on task context}

**Expected Behavior:**
- {Inferred from task descriptions}

**Similar existing:** {If discovery found similar items, list them as reference}

---

### 2. {next_capability}
...