# Llama Lounge Agentic Hackathon: Project Gallery - **Event:** [Llama Lounge Agentic Hackathon](https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon) - **When:** Feb 27 at 9:00 AM – Feb 28 at 10:00 PM (PST) - **Where:** 135 Constitution Dr, 8th floor, Menlo Park - **Hosts:** [Snowflake](https://cerebralvalley.ai/u/snowflake), [Cerebral Valley](https://cerebralvalley.ai/u/events) - **Projects:** 24 (6 placed) - **Page:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery ## Projects ### 1. AgenticMesh2026! I run an agency because people want to work with me on gigs and I find great ones! So this will help me reduce the 4-6 hours of manually reconciliation of my subcontractors to get my invoices out to my prime... With AI! - **Team:** [John Miller](https://cerebralvalley.ai/u/jymillersf) - **GitHub:** https://github.com/jymiller/llama-hack - **Demo video:** https://www.youtube.com/watch?v=QdKfReagWzw - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=1 ### 2. Agent Watchdog AI agents like Cursor, Claude, and Copilot now run shell commands, read files, and even make payments through agentic wallets. They have real-world power without a permission system. A prompt injection or buggy plugin can leak API keys, run dangerous CLI commands, or spend money without the owner’s knowledge. Today the only defense is “don’t give agents access,” which blocks real productivity gains. We built Agent Watchdog to solve this. It uses system-level process monitoring to block risky actions by default, lets owners grant scoped access with passkey authentication for specific tasks, budgets, and time limits, and provides a dashboard to audit every action and payment. Agents stay autonomous within scope, but the owner always keeps control. - **Team:** [Yusuke Muraoka](https://cerebralvalley.ai/u/notsofat), [Isabella kang](https://cerebralvalley.ai/u/kang_dev) - **GitHub:** https://github.com/not-so-fat/agent_watchdog - **Demo video:** https://youtu.be/e3Y2_oifxcU - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=2 ### 3. Autonomyx Four AI agents autonomously handle enterprise support tickets end-to-end. A customer reports false disk faults on a ThunderRack X9000 server. The Intake Agent chats with the customer and creates a ticket in Snowflake. The Diagnosis Agent searches KB articles and bug patterns, scores 73% confidence, and routes to code review. The Code Analyst finds the exact bug in firmware source code and creates a real GitHub Pull Request with the fix. The Resolver tracks SLA, notifies the customer, and closes the ticket. - **Team:** [Sandeep Singh](https://cerebralvalley.ai/u/sandeepsign) - **GitHub:** https://github.com/sandeepsign/autonomyx - **Demo video:** https://youtu.be/DiEHXMozIuQ - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=3 ### 4. Lobster Control a lobster in an immersive 3d environment for completing awesome tasks! Lobster is a 3d console to interact with your lovely OpenClaw bot easily and lazily. - **Team:** [Cal Cu](https://cerebralvalley.ai/u/calcu), [Lily Su](https://cerebralvalley.ai/u/lilyxsu) - **GitHub:** https://github.com/mochiyaki/lobster - **Demo video:** https://github.com/mochiyaki/lobster-office-adventure/blob/main/video.gif - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=4 ### 5. Workflow Guardrail Workflow Guardrail is an agentic system embedded inside creative workflows that prevents costly scope creep before it propagates. A central Orchestrator routes user actions to specialized agents including ScopeWatcher, ChangeTriage, and CostCalculator. The system evaluates changes against project constraints and historical workflow data stored in Snowflake, surfaces real-time deadline and cost impact, and gates message delivery when necessary. This is agentic AI governing outcomes — not just generating responses. - **Team:** [Uma Kelkar](https://cerebralvalley.ai/u/uma) - **GitHub:** https://github.com/paintlikeanengineer/workflow-guardrail - **Demo video:** https://youtu.be/dPrqIS4E0bc - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=5 ### 6. KAIRO: Your AI Chief of Staff Kairo features Ghost Mode for autonomous email/Slack triage with confidence-gated auto-replies and VIP escalation, a Relationship Intelligence graph tracking sentiment, tone shifts, and interaction patterns across all contacts, Commitment Tracking that NLP-extracts promises and scores reliability, Burnout Detection with risk scoring and one-click interventions, Flow State Guardian for deep focus protection, Smart Delegation with AI-ranked candidate matching, Decision Replay for counterfactual "what-if" reasoning, Agent Mesh for multi-user agent coordination with audible negotiations, an Agent Marketplace with live USDC micropayments via Skyfire, a bilingual voice interface (three AI personalities), NLP dashboard commands, and AI-generated weekly analytics reports. Built with CrewAI + Composio OAuth, Snowflake cloud memory, Skyfire payments, LiveKit + Deepgram voice, FastAPI, and Next.js 15. App Link: https://frontend-zeta-lyart-89.vercel.app/ - **Team:** [Gaurav Gupta](https://cerebralvalley.ai/u/gauravright), [Phani Kulkarni](https://cerebralvalley.ai/u/phanikulkarni) - **GitHub:** https://github.com/AIEdtech/Kairo - **Demo video:** https://www.loom.com/share/cb872798752e48608e2b49dedb8e94c5 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=6 ### 7. DeepWorkAgent A persistent background agent that monitors your browser AND entire Mac system in real time, infers what task you're working on from your calendar and activity, and autonomously intervenes when you drift — with precision, not blunt force. It understands that a developer watching a FastAPI tutorial on YouTube is working. It knows the difference between Slack in your project channel vs. Slack in #random. It reads your calendar to know what you should be doing, and escalates gradually rather than blocking immediately. - **Placement:** Finalist - **Team:** [Debbie Viona](https://cerebralvalley.ai/u/dviona), [Shloka Pandya](https://cerebralvalley.ai/u/shloka), [Katherina Nguyen](https://cerebralvalley.ai/u/rkr) - **GitHub:** https://github.com/shloka-812/DeepWorkAgent - **Demo video:** https://youtu.be/Dxw6WnPRSpc - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=15 ### 8. SoliDefy Solopreneurs and one-person businesses (millions in the US alone, contributing trillions to the economy) face massive time sinks and revenue leaks: manually researching leads, crafting proposals, sending invoices, chasing payments, tracking revenue trends, and paying for tools/vendors. This cognitive overload kills momentum—founders spend hours on ops instead of building/growing. Existing tools are fragmented (CRM + email + invoicing + spreadsheets), reactive, and require constant human intervention. In the agentic age, there's no unified, proactive "COO-in-your-pocket" that autonomously handles the full revenue flywheel while staying within budget and improving over time. Our Solution: SoloForge, An autonomous Solopreneur OS Agent (built as a CrewAI-orchestrated multi-agent system) that turns one email or idea into closed-loop revenue—researching leads, drafting personalized proposals, sending invoices, collecting payments, and forecasting growth—all while you focus on your core work. - **Team:** [Greesan Gurumurthy](https://cerebralvalley.ai/u/greesan), [Mahesh Jammalamadaka](https://cerebralvalley.ai/u/jmaheshrao), [Jose Bulatao](https://cerebralvalley.ai/u/bulatao), [Tanusri Jammalamadaka](https://cerebralvalley.ai/u/jtanusri) - **GitHub:** https://github.com/jmtworldwide/soloforge - **Demo video:** https://www.youtube.com/watch?v=kFSy423_xzk - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=7 ### 9. LeadSync LeadSync:Let's keep software teams in sync. AI code tools like Copilot generate code but ignore your team's architecture, standards, and context. Devs ping tech leads for every ticket, losing hours to repeated explanations scattered across Jira, GitHub, and Slack. Multi-agent system that injects team intelligence into workflows. Agents pull live tech lead preferences (editable Google Docs), repo history, and tickets to automate guidance. Ticket Enrichment: Jira webhook → agents enrich tickets with paste-ready prompts (context + rules). Daily Digest: Scans GitHub commits, groups by area, posts Slack summary. Slack Q&A: /leadsync TICKET-123 "Redis or Memcached?" → preference-aligned answer. PR Auto-Description: Enriches PRs with details + validation. Unique: Preferences update instantly via Google Docs. SQLite memory compounds context over time. Agents live in cloud, react to events proactively. - **Team:** [Ramis Hasanli](https://cerebralvalley.ai/u/ramizik), [Hong Doan](https://cerebralvalley.ai/u/hongdnn) - **GitHub:** https://github.com/hongdnn/leadsync - **Demo video:** https://youtu.be/fAU51vf3yuk - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=9 ### 10. Second Soul What It Is Second Soul is a motive-driven, multi-agent personal AI system. It acts as a digital "second self" that monitors your health, finances, and productivity — then takes action on your behalf. Three AI agents (Health, Finance, Productivity) work under a Chief of Staff orchestrator that aligns every decision to your primary life goal. The Problem People are drowning in personal data scattered across apps — sleep trackers, bank accounts, task managers, calendars. No single tool connects the dots. You might be overspending on coffee while your sleep is declining, and no system tells you these are related or suggests what to do about it. Existing AI assistants answer questions but don't act — they can't buy the supplement your health data says you need or alert you on Discord when your spending spikes. - **Team:** [Keshav Dalmia](https://cerebralvalley.ai/u/keshavdalmia10), [Ash Castelino](https://cerebralvalley.ai/u/ash1512) - **GitHub:** https://github.com/Build-for-fun/second_soul - **Demo video:** https://youtu.be/UsbsMwNp5XE - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=16 ### 11. DocRight Context Coding agents like Claude Code, Cursor, and Copilot write great code but hallucinate APIs. Ask one to "build a CrewAI agent with Composio Slack tools" and it uses method names that don't exist. Why? Stale training data, not actual docs. DocAgent fixes this. It indexes doc sites via llms.txt, reasons about which pages are relevant across multiple sources, fetches only those, and generates code using real API patterns. Today: your coding agent guesses, you debug. With DocAgent: it reads the real docs first, your code works. Long term, any AI tool calls our API for accurate context before generating code. Via Skyfire, agents pay per query to learn new APIs. As doc sites adopt llms.txt, DocAgent scales with zero re-engineering. Your coding agent hallucinates APIs. Ours reads the docs first. - **Placement:** Finalist - **Team:** [Rajashekar V](https://cerebralvalley.ai/u/raj) - **GitHub:** https://github.com/rajashekarcs2023/llama-lounge-agentic - **Demo video:** https://youtu.be/omO7fJ0LipE - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=8 ### 12. ToolForge ToolForge is an autonomous AI agent that solves a core limitation of traditional AI: being stuck with only the tools it was given. Instead of a fixed set of integrations, ToolForge starts with nothing and builds what it needs on the fly. Its Planner agent checks existing tools first, then either routes through Composio's 150+ integrations or spins up the Builder agent to write custom code from scratch, test it, and verify it — no human needed. Every tool built gets stored in a Snowflake registry, so the agent compounds knowledge over time. Orchestrated by CrewAI, reasoned over by Llama via Snowflake Cortex, and capable of autonomously paying for APIs through Skyfire — ToolForge is one of the few agents that truly operates end-to-end. - **Placement:** 2nd Place - **Team:** [Abraham Bhatti](https://cerebralvalley.ai/u/abhatti), [Yash Sanghvi](https://cerebralvalley.ai/u/yashy10), [Rishi Shiramshetti](https://cerebralvalley.ai/u/rshirams) - **GitHub:** https://github.com/rshiramss/tool_forge - **Demo video:** https://youtu.be/C-uxlKpvviQ - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=10 ### 13. Chargeback Agent Merchants lose billions to chargebacks every year — not because the transactions were fraudulent, but because building a winning dispute response is slow, manual, and easy to deprioritize. Most merchants never respond at all. Chargeback Agent automates the entire dispute process. When a chargeback arrives, it immediately pulls transaction data, customer history, and IP intelligence from your database, writes a professional dispute narrative with AI, and delivers a complete evidence package to your payment processor — no human required. - **Team:** [Nelson Lai](https://cerebralvalley.ai/u/chineseman) - **GitHub:** https://github.com/chinesepowered/hack-snowflake/ - **Demo video:** https://www.youtube.com/watch?v=qxclmpi3AGU - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=11 ### 14. Vouch AI AI coding agents can work end to end. They pick packages, hit walls, debug, find workarounds. But all that knowledge dies when the session ends. Agents shouldn't have to work in a silo, rediscovering the same problems from scratch. So we built Vouch. It's a CLI tool that runs inside Claude Code that lets agents review packages and learn from other AI agents. When your agent evaluates a package, it checks what other agents already learned. When it's done, it automatically logs its experience back to the network. You install it in the CLI once. Vouch runs in the background. Your coding agent makes decisions based on real-world signal from other agents, not stale training data or marketing docs. Your agent gets smarter and you ship faster. - **Team:** [Girri Palaniyapan](https://cerebralvalley.ai/u/girri), [Muthhukumar Malaiiyyappan](https://cerebralvalley.ai/u/malai) - **GitHub:** https://github.com/gpalrepo/vouch-ai-fe - **Demo video:** https://www.loom.com/share/a8999abcbdc14d3e9443d30f046993a1 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=12 ### 15. OpenSourceOps OpenSourceOps is an AI-powered OSS due diligence platform for companies. It scouts and ranks open source tools, analyzes project health, security, compliance, and operational readiness, and generates context-aware, evidence-backed Go/Conditional Go/No-Go decisions with prioritized next steps. - **Team:** [Siva Girish](https://cerebralvalley.ai/u/sivagirish) - **GitHub:** https://github.com/sivagirish81/OpenSourceOps - **Demo video:** https://www.loom.com/share/96683e9e6b304b488230137602bd6aea - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=13 ### 16. Orq Orq is an autonomous engineering intelligence layer that understands what teams ship and turns repository activity into structured knowledge and real-world execution. It analyzes code changes, generates release notes, coordinates workflows across GitHub, Slack, and Docs, and can research technical candidates by evaluating their public work to assess role fit. Orq replaces manual reporting and fragmented context with coordinated, multi-agent execution embedded directly into modern engineering workflows. - **Team:** [Yug More](https://cerebralvalley.ai/u/yugmore13), [Sean Aminov](https://cerebralvalley.ai/u/SeanAminov) - **GitHub:** https://github.com/SeanAminov/Orq - **Demo video:** https://youtu.be/n_bISnFWzJE?si=BLhUYgsdGNGBAGeV - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=14 ### 17. GovCortexAi GovCortexAI V2 is a anakytics agent that saves it learning as python notebooks which get approved by data team. It uses CrewAI for agent orchestration, FastAPI for backend APIs/streaming events, Streamlit for the UI, Composio + Snowflake for schema/query execution, and controlled SQL/Python “notebook memory” cells to enforce approved analytics logic and traceability. Skyfire for data buying for complex issues - **Team:** [vivek shukla](https://cerebralvalley.ai/u/weebhek) - **GitHub:** https://github.com/vivek100/Govcortexai - **Demo video:** https://youtu.be/RY55hlkeEdI - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=17 ### 18. AgentOps AgentOps uses a CrewAI multi-agent pipeline to autonomously detect, diagnose, and fix ML production issues. Three agents work in sequence: - Monitor pulls metrics and feature drift from Snowflake, detecting anomalies against baselines - Investigator performs root cause analysis using RAG over runbooks and past incidents - Remediator executes fixes via Composio integrations with Slack and GitHub Each agent builds on the previous agent's output—mimicking how an on-call engineer triages, diagnoses, and resolves issues. A React dashboard provides human-in-the-loop oversight: approving actions, monitoring agent progress via live streaming, and initiating investigations. - **Team:** [Felix Amoruwa](https://cerebralvalley.ai/u/felix_amoruwa) - **GitHub:** https://github.com/ovai-felix/agentops - **Demo video:** https://youtu.be/JIIJSReAOAg - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=22 ### 19. BarackOllama Our project is a personalized, adaptive resume generation agent that automates the job application workflow. It uses Bright Data to scrape LinkedIn and GitHub data, leverages CrewAI for multi-agent reasoning, and integrates external tools through Composio to analyze job descriptions and perform intelligent, personalized keyword alignment with the user’s profile. The system then generates a dynamic HTML resume and converts it into a ready-to-download PDF, transforming structured data and agent-based reasoning into a real-world, customized application artifact. - **Team:** [Ankita Sethi](https://cerebralvalley.ai/u/ankitasethi), [Shreyas Yadav](https://cerebralvalley.ai/u/shri_), [Himavanth Karpurapu](https://cerebralvalley.ai/u/hima), [Csaba Toth](https://cerebralvalley.ai/u/tocsa) - **GitHub:** https://github.com/LlamaLoungeSnowflake/BarackOllama - **Demo video:** https://youtu.be/-JnvK9SV4N0 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=18 ### 20. YesChef YesChef is a voice agent that listens to multi-person meeting and execute agentic actions in real time - **Team:** [William Gao](https://cerebralvalley.ai/u/williamsf), [Zurab Kachukhashvili](https://cerebralvalley.ai/u/zurakach), [Tommy Zhao](https://cerebralvalley.ai/u/tmyzhao) - **GitHub:** https://github.com/dcstup/meetingagent - **Demo video:** https://www.loom.com/share/c8de0327079348df9cd810ada84ea139 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=19 ### 21. spartans Team Spartan | Shubham Kothiya, Prem Shah, Kushal Adhyaru, Mann Nada As enterprises deploy dozens of AI agents and integrate rapidly evolving open-source repos, production failures cascade across interconnected services faster than humans can respond. IncidentDNA detects anomalies and runs a multi-agent AI pipeline to investigate, validate, and resolve incidents autonomously — in under 2 minutes, no human required. Three CrewAI agents powered by Claude Sonnet 4.5 (Snowflake Cortex) collaborate via A2A communication: Agent 1 classifies severity and maps blast radius, Agent 2 investigates root cause using runbooks and past incidents, Agent 5 adversarially debates the hypothesis. A threshold engine auto-resolves or escalates to humans. Actions are executed via Composio — Block Kit Slack alerts with approve buttons and detailed GitHub issues with fix options, all idempotent and fully auditable in Snowflake. Stack: Snowflake, CrewAI, Composio, FastAPI, React - **Team:** [Shubham Jaysukhbhai Kothiya](https://cerebralvalley.ai/u/shubh07), [Kushal Atulbhai Adhyaru](https://cerebralvalley.ai/u/kushal511), [Mann Nada](https://cerebralvalley.ai/u/mannnada), [Prem Shah](https://cerebralvalley.ai/u/premshah) - **GitHub:** https://github.com/theshubh007/IncidentDNA - **Demo video:** https://youtu.be/bGARkL6evDM - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=20 ### 22. Logentic Logentic is an AI-powered freight visibility platform for intermodal rail brokers.Instead of relying on phone calls and scattered emails, carriers submit shipment updates through a simple web form. Logentic then instantly notifies shippers by email, SMS, or voice, and provides a self-service tracking page for real-time status. Built on a Snowflake-native data layer with Cortex LLM, CrewAI + Composio automation, and Skyfire billing settlement, Logentic integrates with existing broker workflows. The result: faster communication, real-time visibility, and far less manual coordination. - **Placement:** 3rd Place - **Team:** [Sungman Cho](https://cerebralvalley.ai/u/sungman), [Jaffar Keikei](https://cerebralvalley.ai/u/jaffarkeikei), [James Nicholson](https://cerebralvalley.ai/u/nicholsonjf) - **GitHub:** https://github.com/sungmanch/Logentic - **Demo video:** https://youtu.be/QpefUkrXbwk - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=21 ### 23. OrchestrAI Troubadour AI is an autonomous music orchestration agent that observes real-time events, reasons about emotional context, and dynamically directs adaptive soundtracks and narration using CrewAI, Snowflake, and ElevenLabs. Today, game audio tools are static and rule-based, requiring manual scripting; no autonomous agent exists that can interpret live environments and orchestrate music intelligently. This fills a major market gap across games, VR, films, simulations, and interactive media, where emotional adaptation is critical but currently manual and inflexible. In the future, we will enable agent-to-agent payments via Skyfire, allowing game engines and digital environments to automatically discover, license, and pay music agents in real time, creating a fully autonomous audio economy. - **Placement:** 1st Place - **Team:** [Jidhnyasaa Mahajan](https://cerebralvalley.ai/u/Jidhnyasaa), [Mihai Bulic](https://cerebralvalley.ai/u/mihai), [Jake Schwartz](https://cerebralvalley.ai/u/jakesc), [Victoria Duan](https://cerebralvalley.ai/u/jq_victoria_d) - **GitHub:** https://github.com/JakeSc/Snowflakathon/tree/master - **Demo video:** https://www.loom.com/share/b3fdb5659964492ea9e6181a94b4cbb0 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=23 ### 24. Draper Draper is a full-stack agentic team platform that autonomously handles sales, marketing, and GTM operations. It deploys AI agents that prospect leads, make voice calls, draft personalized emails, schedule meetings, sync CRMs, and even hire humans for in-person meetings -- all from a single command console. - **Placement:** Finalist - **Team:** [Yeabsira Mulugeta](https://cerebralvalley.ai/u/YAB) - **GitHub:** https://github.com/xenn0010/Draper - **Demo video:** https://www.loom.com/share/ee4c54266ea54c919eb764c5cd532211 - **Project:** https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery?project=24 --- Markdown version of https://cerebralvalley.ai/e/llama-lounge-agentic-hackathon/hackathon/gallery. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt