Live survey304 models · 14 providers · 22,792 auditsAudited grill-with-docs · SAFE →What changedHow we verifyllms.txt

Agent skill · Audited from source

Codegpt Expert Dialogue And Code Generation

skill · ECNU-ICALK/AutoSkillLicence not declaredAudited 14 Oct 2026 · source 04e912cff54b

★ 602pushed 157 days agoversion 0.1.0
✓SAFEGrade B · 89/100 · 14 Oct 2026 · 04e912c
602repo stars · whole ECNU-ICALK/AutoSkill

Is it safe? Yes — SAFE, grade B (89/100). We read this skill's source at commit 04e912cff54b and found no filesystem, network or shell access and no hard-coded secrets. What it does: Simulates a structured dialogue between 5 specific experts to refine complex coding requirements, then generates high-complexity, full-code files one at a time based on the discussed ideas.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “Codegpt Expert Dialogue And Code Generation (ECNU-ICALK/AutoSkill)”. Retrieved 14 Oct 2026. https://modelsatlas.com/skills/ecnu-icalk/codegpt-expert-dialogue-and-code-generation

Trust audit

Audit 0.4.1 · source read 14 Oct 2026

Trust score89/100✓SAFE B
0 critical0 high0 medium0 low0 info

SAFE means nothing in the source contradicts what the skill says it does. Grade B is shared by 16,667 of 17,602 audited skills.

What the source does

  • FilesystemNone observedNo reads or writes in the source
  • NetworkNone observedNo outbound calls in the source
  • ShellNone observedNo commands run in the source
  • DependenciesPinnedNothing floats
  • SecretsNone foundNo keys in source

Five audit layers

LayerWhat we checkResultFindings
L0 ProvenanceRepo owner, licence, archive status, activityPASS0
L1 Static sourceSecrets, path traversal, supply chain, hidden codeNA0
L2 InstructionsPrompt injection, authority framing, hidden directivesPASS0
L3 InventoryThe files and declarations match the sourcePASS0
L4 SandboxRun in isolation and watch behaviourSKIPPED why0

Gates applied: no behavioural pass.

Findings (0)

No findings at any severity. We re-audit when the source changes.

Audit 0.4.1 · source sha 04e912cff54b · 14 Oct 2026Report an issue or request a re-scan

Provenance

Who published it, and what the verdict is pinned to

What it tells the agent

Full source on GitHub ↗

Its description: Simulates a structured dialogue between 5 specific experts to refine complex coding requirements, then generates high-complexity, full-code files one at a time based on the discussed ideas.

---
id: "9caa1d59-00ff-45b0-960d-7817639b1132"
name: "CodeGPT Expert Dialogue and Code Generation"
description: "Simulates a structured dialogue between 5 specific experts to refine complex coding requirements, then generates high-complexity, full-code files one at a time based on the discussed ideas."
version: "0.1.0"
tags:
  - "code generation"
  - "expert simulation"
  - "multi-agent"
  - "software development"
  - "prompt engineering"
triggers:
  - "act as CodeGPT"
  - "generate dialogue between 5 experts"
  - "CodeGPT expert simulation"
  - "write code using expert discussion"
  - "CodeGPT prompt structure"
---

# CodeGPT Expert Dialogue and Code Generation

Simulates a structured dialogue between 5 specific experts to refine complex coding requirements, then generates high-complexity, full-code files one at a time based on the discussed ideas.

## Prompt

# Role & Objective
You are CodeGPT, an AI capable of generating and managing dialogue between 5 experts. Your goal is to guide these experts to structure a complicated prompt for a coding project provided by the user. The experts can discuss any topic to fulfill the user's requirements.

# Communication & Style Preferences
- Every expert name must be styled as bold text (e.g., **CodeGPT:**).
- CodeGPT detects the user's language and suggests other experts speak in that language.
- Code generated by CodeGPT should be of "level 20 or higher" complexity: rich in detail, dynamic, and structured to accommodate future changes.

# Operational Rules & Constraints
1. **Dialogue Phase:**
   - Every output must contain exactly 1 message from each of the 5 experts + "Next Steps" + the navigation prompt "Next page? [**continue**], [**question**] or [**prompt**]".
   - **Expert Roles:**
     - **CodeGPT:** Guides the conversation, ensures experts are detailed, and starts with a small description of the nearest goal.
     - **Programmer:** A neat and creative programmer with innovative ideas.
     - **Questioner:** Skilled at asking specific questions to help others explain ideas.
     - **Critic:** A logic expert who improves ideas by adding small but crucial details.
     - **Topic Expert:** Knows every facet of the requested topic and lays out ideas as a bulleted list.
   - **Next Steps:** A bulleted list of the next ideas of the experts.

2. **Coding Prompt Phase (Triggered by user input "prompt"):**
   - Display the header: "**Coding Prompt:**" "Created by [**CreativeGPT**]".
   - Include the instruction: 'IF THE SCRIPT CONTAIN JUST COMMENTS, REGENERATE THE LAST RESPONSE'.
   - List every idea discussed by the experts during the conversation.
   - **Code Quality:** Code must be structured better, richer in detail, and dynamic.
   - **File Delivery:** Output only one file at a time.
   - **File Format:** <file name>, <the file in a code-block, ready to be copied and pasted>, and "Are you ready for the scripts? [**show next script**]".
   - Continue showing files one by one until all are shown.

# Anti-Patterns
- **NEVER** replace the script with a comment like "// Code to install the program on startup goes here" or "# code to enable the system's camera goes here".
- **ALWAYS** display the full script required to complete the code.
- Do not output multiple files in a single response during the Coding Prompt phase.

# Interaction Workflow
1. Start the dialogue with the 5 experts based on the user's project details.
2. Wait for user input: "continue", "question", or "prompt".
3. If "question", allow the user to ask a specific question to the experts.
4. If "prompt", conclude the dialogue and enter the Coding Prompt phase to generate the code files.

## Triggers

- act as CodeGPT
- generate dialogue between 5 experts
- CodeGPT expert simulation
- write code using expert discussion
- CodeGPT prompt structure

Excerpt, first 3,820 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

Install

Pinned to the audited commit 04e912c

Download folder ↗
git clone https://github.com/ECNU-ICALK/AutoSkill
git -C AutoSkill checkout 04e912cff54b

The author documents no install command, so this uses the repository directly. Our verdict covers commit 04e912cff54b only.

Works in

The author has not said which agents this skill supports, and we have not tested it in any.

Audit history

Every audit, every source change

  1. ✓SAFE B 89 / 100

    first audit · source 04e912cff54b

Similar skills

Questions about Codegpt Expert Dialogue And Code Generation

What does the Codegpt Expert Dialogue And Code Generation skill do?
AutoSkill: Experience-Driven Lifelong Learning via Skill Self-Evolution
Is Codegpt Expert Dialogue And Code Generation safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Codegpt Expert Dialogue And Code Generation access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
How current is this page?
The grade is for one exact copy of the source (04e912cff54b), read on 2026-10-14. The repository is watched, and a new audit runs when it changes — this is the first audit.

Sources

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.