Skip to content
Back to skills

Phase2 DailyChat

ASecurity

Use this skill when you only have the index XML file (the high-level map) and you need to see the full source code of specific files to complete the user's request.

  • 4 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 2, 2026
ai-agentsapi

Works with

  • api

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 2, 2026

npx -y skills add arpanvgm/ai-bridge --skill Phase2-DailyChat --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Phase2 DailyChat?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Phase2 DailyChat
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/arpanvgm-phase2-dailychat/badge)](https://www.skillsdirectory.com/skills/arpanvgm-phase2-dailychat)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: ai-request-skill
description: >
  Use this skill when you only have the index XML file (the high-level map) and you need to see the full source code of specific files to complete the user's request.
---

# AI Bridge Request Skill

When you are working from the index XML file map, you do not have the full source code for the files. To perform complex coding tasks, you must ask the system to provide the full contents of the specific files you need.

## RESPONSE STRUCTURE

To request file contents, you MUST output a strict XML block. The requested files will be read and provided to you in the next prompt.

```xml
<ai-request>
  <file path="Relative/Path/To/File1.cs" />
  <file path="Relative/Path/To/File2.cs" />
</ai-request>
```

## RULES FOR REQUESTING FILES

1. **Exact Paths:** The `path` attribute must exactly match the `path` attribute found in the index XML file.
2. **Batch Requests:** Request all the files you think you will need for the task in a single `<ai-request>` block to save time.
3. **No Code Output Yet:** Do not attempt to guess the code or write an `<ai-response>` patch in the same message as an `<ai-request>`. Wait for the file contents to be provided before you write any code modifications.
4. **Markdown Formatting:** Wrap the XML block in a standard markdown ````xml ```` block.

## REPLY FORMAT (what you receive)

The reply wraps requested files in `<module>` blocks, grouped by module:

```
<module name="WebApi" files="1">
<file path="WebApi/Controllers/UserController.cs" lines="62">
// full source code
</file>
</module>

<module name="Shared" files="1">
<file path="Shared/Models/UserProfile.cs" lines="14">
// full source code
</file>
</module>
```

If a requested file and its content are missing from the reply (e.g. it was renamed or
deleted since the index was generated), do not guess its content — tell the user the
file could not be found and ask how to proceed.

## WORKFLOW EXAMPLE

**User:** "Can you add a new API endpoint for fetching user profiles?"
*(User has only provided the index XML file)*

**AI:**
I need to see the current Controllers and the User model to implement this.

```xml
<ai-request>
  <file path="WebApi/Controllers/UserController.cs" />
  <file path="Shared/Models/UserProfile.cs" />
</ai-request>
```

**Reply (What you receive in next prompt):**

```
<module name="WebApi" files="1">
<file path="WebApi/Controllers/UserController.cs" lines="62">
// full source code
</file>
</module>

<module name="Shared" files="1">
<file path="Shared/Models/UserProfile.cs" lines="14">
// full source code
</file>
</module>
```

**AI:**
Now has full source for both files, and proceeds to write the `<ai-response>` as per
`ai-response-skill.md` — no further `<ai-request>` needed for this task unless
additional files turn out to be necessary.

Files in this skill

  • ai-index-system-prompt.md4.2 KB
  • ai-request-skill.md2.8 KB
  • ai-response-skill.md9.9 KB
  • ai-update-index-skill.md4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.