Add agent hooks for post functionality

[danieldiaz-qaa]

This pull request adds a new unpublished post, "Your agent follows instructions. Until it doesn't.", which introduces and explains Agent Hooks in VS Code. The post provides a comprehensive overview of how Agent Hooks work, their importance compared to instructions and skills, practical usage examples, safety considerations, and integration with other customization features.

Key additions and themes:

New Content: Agent Hooks Overview

• Adds a detailed markdown post (2.agent-hooks.md) explaining what Agent Hooks are, how they differ from instructions and skills, and why they matter for deterministic automation in agent workflows.

Practical Guidance and Examples

• Includes practical code samples for defining and using hooks, such as auto-formatting files, blocking destructive commands, and injecting runtime context into agent sessions.
• Explains the eight available hook events in VS Code, their lifecycle points, and how to use them effectively.

Safety and Troubleshooting

• Highlights important safety considerations, including script permissions, validation of input, and preventing agents from editing hook scripts.
• Provides troubleshooting tips for when hooks do not behave as expected, focusing on exit codes and the VS Code output channel.

Cross-Platform and Integration Notes

• Notes compatibility and differences with Claude Code hooks, and how VS Code supports reading from Claude-style configuration files.

[Copilot]

Pull request overview

This PR adds a new unpublished markdown post introducing VS Code Agent Hooks and how they fit into the broader Copilot customization workflow.

Changes:

• Adds a full Agent Hooks explainer post with lifecycle event descriptions.
• Includes hook configuration examples for formatting, blocking commands, and injecting context.
• Adds safety, troubleshooting, Claude compatibility, and customization-stack guidance.

Comments suppressed due to low confidence (1)

posts/danieldiaz-qaa/2.agent-hooks.md:210

• This section states that VS Code hook inputs use camelCase, but the earlier PreToolUse example and parser use tool_name / tool_input in snake_case. One of these examples is wrong; align the payload shape and parser throughout the post so readers can copy a working hook.

- **Tool input property names**: Claude Code uses `snake_case` (e.g. `tool_input.file_path`), VS Code uses `camelCase` (e.g. `tool_input.filePath`)
- **Tool names**: different. Claude uses `Write`/`Edit`, VS Code uses `create_file`/`replace_string_in_file`

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Dry-run output:

[UPDATED] [PUBLISHED] Stop getting generic output from Copilot. Teach it y… 
[CREATED] [PUBLISHED] Your agent follows instructions. Until it doesn't.