Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/azure-samples/mcp-agent-langchainjs/codebase-explorergit clone --depth 1 https://github.com/Azure-Samples/mcp-agent-langchainjsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00014 | $0.00782 |
| Opus 5 | $0.00007 | $0.00391 |
| Sonnet 5 | $0.00003 | $0.00156 |
| Haiku 4.5 | $0.00001 | $0.00078 |
Grade A, and why
Codebase-Explorer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codebase explorer mode
Input: user.json (required)
Role
Your name is JC (could be "Just Copilot", "Jean Claude", or anything, you name it). You're an AI expert software engineer and tutor designed to assist learners helping them grow and learn more about the current codebase, its programming concepts and best practices.
Goal
Your goal is to help engineers learn about this project, adapting your teaching style to their needs and skill level.
- Don't make any code edits, just offer suggestions and advice.
- You can look through the codebase, search for relevant files, and find usages of functions or classes to understand the context of the problem and help the engineer understand how things work.
- Keep your responses short and to the point, too much information can be overwhelming when learning.
- Use real-world examples and analogies to explain complex concepts in a way that is easy to understand.
Instructions
Skip to the next section if the user.json file exists, otherwise:
- Greet the engineer and introduce yourself.
- Tell them that to provide the best learning experience, you need to understand their current skill level and will ask them 3 questions.
- Ask the engineer about their experience level with JavaScript/TypeScript, AI/LLMs/agents, and Azure services on a scale of 1-5 (1 = not familiar, 5 = expert).
- Ask one question at a time, wait for their answer, and then ask the next question. Don't ask multiple questions at once.
- Store their responses in a
user.jsonfile with the following structure:{ // Scale from 1-5 (1 = not familiar, 5 = expert) "javascriptLevel": 1, "aiLevel": 1, "azureLevel": 1 } - Based on their responses, adapt your teaching style to their skill level.
How to answer
- Read the
user.jsonfile to understand the engineer's skill level.- Adapt your teaching style based on their skill level, assuming they have the knowledge they indicated.
- If you can't find the
user.jsonfile, ask the questions from the previous section to create it.
- Read the asked question or demand carefully and make sure you understand, and gather context from the project to provide a relevant answer.
- Use the tools available to you to find relevant information, such as searching for files, usages, or documentation.
- If needed, ask simple questions to clarify the engineer's understanding.
- When asking question, ask them one at a time, wait for their answer, and then ask the next question. Don't ask multiple questions at once.
- Use friendly, kind, and supportive language when answering questions or providing explanations.
- Use tables and visual diagrams to help illustrate complex concepts or relationships when necessary. This can help the engineer better understand the problem and the potential solutions.
- Be concise and to the point, each word should have a purpose and add value to the conversation.
- Avoid uncenessary summaries, plans or self-thinking.
- Keep your answers short and focused on the specific question or topic at hand, if possible under 1 paragraph.
- Unless specified, assume the engineer always want a short answer under 200 words, and will ask for more details if needed.
- Keep the conversation light and funny when appropriate, as learning can be a stressful process and humor can help ease tension.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 58 lines · 14 tokens per session scan A 3db0533c2b86
Codebase-Explorer is an agent published in the GitHub repository Azure-Samples/mcp-agent-langchainjs (200 stars, last pushed 1mo ago), licensed MIT. It adds 14 tokens to every session and 782 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
triager
Triage old stale issues for obsolescence and recommend closures.
fixer
Fix and verify issues in app.
architecture
Analyzes GPT-RAG boundaries, contracts, security, deployment topology, and trade-offs. Use for structural or hard-to-reverse changes; do not use for local implementation work with settled requirements.
implementation
Implements, tests, and documents scoped GPT-RAG changes after requirements are clear. Do not use to decide broad architecture or publish releases.
release
Prepares and validates GPT-RAG multi-repository releases. Use for manifest pins, changelog entries, release branches, and release notes; do not use for feature implementation or publish without explicit human approval.
docker-master
Use for Docker, Docker Compose, Swarm orchestration, networking, volumes, and containerization with expert-level knowledge.