From c7450da26d9da97fb41c8db913f4a143ebb81d14 Mon Sep 17 00:00:00 2001 From: GAP Promoter Date: Wed, 27 May 2026 14:40:32 +0000 Subject: [PATCH 1/2] Add stepanogil/autonomous-hr-chatbot to the registry --- .../README.md | 54 +++++++++++++++++++ .../metadata.json | 14 +++++ 2 files changed, 68 insertions(+) create mode 100644 agents/stepanogil__autonomous-hr-chatbot/README.md create mode 100644 agents/stepanogil__autonomous-hr-chatbot/metadata.json diff --git a/agents/stepanogil__autonomous-hr-chatbot/README.md b/agents/stepanogil__autonomous-hr-chatbot/README.md new file mode 100644 index 0000000..3d0fddb --- /dev/null +++ b/agents/stepanogil__autonomous-hr-chatbot/README.md @@ -0,0 +1,54 @@ +# Autonomous HR Chatbot + +An autonomous HR assistant agent that answers employee queries using a suite of +specialized tools — all powered by OpenAI's language models. + +## What it does + +Employees can ask natural-language questions about HR policies and their personal +records, and the agent autonomously selects the right tool to answer: + +| Query type | Tool used | +|---|---| +| "What's the policy on unused vacation leave?" | **Timekeeping Policies** — semantic search over the official HR policy document (Pinecone / OpenAI file-search vector store) | +| "How many sick days do I have left?" | **Employee Data** — Python/pandas REPL querying a structured employee dataset | +| "If I work 3 extra hours each day for 2 weeks, how many overtime hours is that?" | **Calculator** — LangChain LLMMathChain / arithmetic | + +## Key capabilities + +- **Policy retrieval** via RAG: embeddings generated with `text-embedding-ada-002` and stored in Pinecone; v2 uses OpenAI's built-in `file_search` tool. +- **Employee data analysis** via a Python AST REPL with access to a pandas dataframe of employee records. +- **Arithmetic** via LangChain's LLMMathChain (v1) or model reasoning (v2). +- **Streaming front-end** built with Streamlit, including live reasoning-summary display in v2. +- Two implementations included: `v1` (LangChain + Pinecone, 2023) and `v2` (OpenAI Responses API + GPT-5, 2026). + +## Tech stack + +- **LLM**: GPT-3.5-turbo (v1) / GPT-5 with reasoning summaries (v2) +- **Orchestration**: LangChain agents (v1) / OpenAI Responses API agent loop (v2) +- **Vector store**: Pinecone (v1) / OpenAI file-search vector store (v2) +- **Front-end**: Streamlit +- **Data**: pandas CSV employee dataset + plain-text HR policy document + +## Example usage + +``` +User: What is the policy on unused vacation leave? +Agent: [searches Timekeeping Policies] According to the policy, unused vacation + leave of up to 10 days may be carried over to the following year... + +User: How many sick leaves do I have left? +Agent: [queries Employee Data] You have 7 sick leave days remaining. +``` + +## Getting started + +See the [README](https://github.com/stepanogil/autonomous-hr-chatbot#how-to-use-this-repo) +for setup instructions. You will need OpenAI and Pinecone API keys (v1) or just +an OpenAI key with a vector store set up (v2). + +## Links + +- **Repo**: https://github.com/stepanogil/autonomous-hr-chatbot +- **Blog post**: https://medium.com/@stephen.bonifacio/creating-a-mostly-autonomous-hr-assistant-with-chatgpt-and-langchains-agents-and-tools-1cdda0aa70ef +- **Video demo**: https://www.youtube.com/watch?v=id7XRcEIBvg diff --git a/agents/stepanogil__autonomous-hr-chatbot/metadata.json b/agents/stepanogil__autonomous-hr-chatbot/metadata.json new file mode 100644 index 0000000..8ea0e4f --- /dev/null +++ b/agents/stepanogil__autonomous-hr-chatbot/metadata.json @@ -0,0 +1,14 @@ +{ + "name": "autonomous-hr-chatbot", + "author": "stepanogil", + "description": "Autonomous HR assistant that answers employee queries using policy retrieval, employee data analysis, and a calculator — powered by GPT and LangChain.", + "repository": "https://github.com/stepanogil/autonomous-hr-chatbot", + "version": "1.0.0", + "category": "customer-support", + "tags": ["hr", "chatbot", "langchain", "openai", "rag", "pinecone", "streamlit", "employee-tools", "policy-retrieval", "autonomous-agent"], + "license": "MIT", + "model": "openai:gpt-4o", + "adapters": ["openai", "system-prompt"], + "icon": false, + "banner": false +} From e7e320ff661653420f0a078d9bb4f25dbcb88807 Mon Sep 17 00:00:00 2001 From: GAP Promoter Date: Sun, 31 May 2026 13:01:01 +0000 Subject: [PATCH 2/2] =?UTF-8?q?Refine=20stepanogil/autonomous-hr-chatbot?= =?UTF-8?q?=20registry=20entry=20=E2=80=94=20improved=20metadata=20+=20pat?= =?UTF-8?q?h=20field?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- agents/stepanogil__autonomous-hr-chatbot/metadata.json | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/agents/stepanogil__autonomous-hr-chatbot/metadata.json b/agents/stepanogil__autonomous-hr-chatbot/metadata.json index 8ea0e4f..414a880 100644 --- a/agents/stepanogil__autonomous-hr-chatbot/metadata.json +++ b/agents/stepanogil__autonomous-hr-chatbot/metadata.json @@ -1,11 +1,12 @@ { "name": "autonomous-hr-chatbot", "author": "stepanogil", - "description": "Autonomous HR assistant that answers employee queries using policy retrieval, employee data analysis, and a calculator — powered by GPT and LangChain.", + "description": "Autonomous HR assistant answering employee queries on leave balances, policies, and org data using RAG, pandas, and a calculator — built with LangChain and OpenAI.", "repository": "https://github.com/stepanogil/autonomous-hr-chatbot", + "path": "", "version": "1.0.0", "category": "customer-support", - "tags": ["hr", "chatbot", "langchain", "openai", "rag", "pinecone", "streamlit", "employee-tools", "policy-retrieval", "autonomous-agent"], + "tags": ["hr", "chatbot", "langchain", "openai", "rag", "pinecone", "streamlit", "autonomous-agent", "employee-tools", "policy-retrieval"], "license": "MIT", "model": "openai:gpt-4o", "adapters": ["openai", "system-prompt"],