How sync-agent-rules.sh Regenerates Platform-Specific Instruction Files

The sync-agent-rules.sh script reads AGENTS.md, resolves Claude-style @file imports, and writes the expanded content to multiple agent-specific configuration files across your repository.

In the JCodesMore/ai-website-cloner-template repository, maintaining consistent AI-agent instructions across different platforms is handled by scripts/sync-agent-rules.sh. This Bash utility eliminates manual duplication by treating AGENTS.md as the single source of truth and automatically regenerating platform-specific instruction files for GitHub Copilot, Cline, Continue, and Amazon Q Developer.

The Regeneration Pipeline

The script operates through a four-phase pipeline that transforms the canonical AGENTS.md into various agent-readable formats.

1. Locating the Source File

The script determines the repository root and sets SOURCE=$REPO_ROOT/AGENTS.md as the canonical instruction file. Safety checks (lines 23-32) abort execution if this source file is missing, preventing incomplete or broken updates.

2. Resolving Claude-Style Imports

Between lines 33-50, the resolve_imports function processes AGENTS.md for lines beginning with @. Each @path reference is replaced with the actual content of the referenced file, producing a fully expanded document stored in RESOLVED_CONTENT. This allows modular rule organization while maintaining a single deployable configuration.

3. Generating the Auto-Generated Header

Lines 55-56 define a standard HEADER variable containing an "AUTO-GENERATED" comment. This header is prepended to every output file, warning developers that the file is derived from AGENTS.md and should not be edited directly.

4. Writing Platform-Specific Files

The write_file helper (lines 58-65) creates target directories if needed and writes the header plus resolved content. Lines 69-86 invoke this helper to generate four distinct configuration files.

Platform-Specific Output Locations

The script generates the following platform-specific instruction files:

• GitHub Copilot Chat: Writes to .github/copilot-instructions.md as a direct copy of RESOLVED_CONTENT.
• Cline / Roo Code: Outputs to .clinerules with identical content.
• Continue: Creates .continue/rules/project.md with YAML front-matter prepended (lines 75-81) followed by the resolved content.
• Amazon Q Developer: Generates .amazonq/rules/project.md as a direct copy.

Running the Regeneration

Execute the script from the repository root to regenerate all platform-specific instruction files:

bash scripts/sync-agent-rules.sh

Expected output:

Syncing agent rules from AGENTS.md...
  ✓ .github/copilot-instructions.md
  ✓ .clinerules
  ✓ .continue/rules/project.md
  ✓ .amazonq/rules/project.md

Done. Generated files are committed to the repo but sourced from AGENTS.md.
Edit AGENTS.md, then re-run this script to update all agent configs.

Summary

• scripts/sync-agent-rules.sh acts as the central orchestrator for AI-agent configuration, reading from AGENTS.md and propagating changes to multiple platforms.
• The resolve_imports function (lines 33-50) handles modular file includes using the @path syntax before generating final outputs.
• Four platforms receive automatic updates: GitHub Copilot (.github/copilot-instructions.md), Cline (.clinerules), Continue (.continue/rules/project.md with YAML front-matter), and Amazon Q (.amazonq/rules/project.md).
• Each generated file contains an AUTO-GENERATED header (defined at lines 55-56) preventing manual edits that would be overwritten.

Frequently Asked Questions

What happens if AGENTS.md is missing when I run the script?

The script performs safety checks at lines 23-32 and aborts immediately if AGENTS.md cannot be found at the repository root. This prevents the generation of empty or corrupted instruction files that would break your AI agent configurations.

Can I modify the generated files directly?

No. The script inserts an explicit "AUTO-GENERATED" warning header into each file to indicate that direct edits will be lost. All changes should be made to AGENTS.md (or files referenced via @ imports) followed by re-running bash scripts/sync-agent-rules.sh.

How do I add support for a new AI agent platform?

Extend the script by adding another write_file call after line 86, specifying your target path and any platform-specific preprocessing (such as the YAML front-matter required by Continue). Follow the existing pattern used for Copilot or Amazon Q if your agent accepts raw markdown.

Why does Continue require special YAML front-matter?

According to lines 75-81 in the source, Continue reads project rules from .continue/rules/project.md but requires a specific YAML metadata block to recognize the file as a rule configuration. The script automatically prepends this block before the resolved content, unlike the direct copies made for other platforms.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}