# Slack Projects at AI Tinkerers > Canonical HTML: https://aitinkerers.org/technologies/slack > Markdown URL: https://aitinkerers.org/technologies/slack.md > Technology record last updated: 2026-09-18T15:13:46Z > Generated: 2026-09-22T14:38:53Z Slack is the digital operating system for work: a real-time messaging and collaboration platform that centralizes communication and tools. Slack is the central hub for team operations, replacing slow email chains with focused, real-time communication. Work is organized into channels (dedicated spaces for projects or teams), direct messages, and threads, ensuring all context is searchable and accessible. It integrates over 2,600 enterprise-ready apps, from Google Drive to Salesforce, bringing data and processes directly into the conversation flow. This structure allows teams to move faster, manage projects efficiently, and keep all personnel (internal and external) aligned on key objectives. - Official technology site: https://slack.com - Public AI Tinkerers demos and talks: 15 - Result page: 1 of 1 ## Recent Public Talks and Demos ### [Slackclaw - OpenClaw for a whole community, on a budget!](https://columbus.aitinkerers.org/talks/rsvp_dJj2d0lQxZ0) Bringing OpenClaw like capabilities to an entire community inside a single slackbot, where everyone gets their own sandboxed workspace, personal memory, and access to scheduling and custom tools. - Event context: AI Tinkerers - Columbus March Meetup β€” 2026-03-02 β€” Columbus - Public talk page: https://columbus.aitinkerers.org/talks/rsvp_dJj2d0lQxZ0 ### [Tricks for building + hosting personal AI systems](https://san-diego.aitinkerers.org/talks/rsvp_i-fmBqL4sFo) Best practices I've learned for building and hosting your own personal AI agents - Event context: AI Tinkerers x Google AI Meetup, San Diego β€” 2025-11-21 β€” San Diego - Public talk page: https://san-diego.aitinkerers.org/talks/rsvp_i-fmBqL4sFo ### [Proactive AI Teammate that evolves with your team](https://nyc.aitinkerers.org/talks/rsvp_Kea85Vzg8pI) We showcase how Teamate, a proactive AI assistant that's designed to work natively with a team, can chime in at the right moment to provide your team with just the right context. We also showcase how it can carry out tasks on behalf of the team, and how it can recall relevant facts so your team can eventually replace your CRM with it. - Event context: Auth0 x AI Tinkerers: Fall Social Demo Night 🍁 ft. Hex β€” 2025-11-06 β€” New York City - Public talk page: https://nyc.aitinkerers.org/talks/rsvp_Kea85Vzg8pI ### [Teamate - Your Proactive AI Teammate in Slack](https://seattle.aitinkerers.org/talks/rsvp_PxtNzvinEgs) I’m building a Slack AI app that introduces a new paradigm of human–AI collaboration: many-to-one. While existing AI systems focus on 1-to-1 (ChatGPT, Claude) or 1-to-many (multi-agent frameworks), this app enables N-to-1 collaborationβ€”where multiple people interact with a single AI in a shared space. Living directly inside Slack channels, the AI listens to group conversations, stepping in at the right moments with suggestions, insights and actions. It acts as a true teammate, creating a richer and more natural way for groups to collaborate with AI in real time. - Event context: Summer on the Lake - August Tinkerers Meetup! (Demos, Food, Friends) β€” 2025-08-28 β€” Seattle - Public talk page: https://seattle.aitinkerers.org/talks/rsvp_PxtNzvinEgs ### [Automate - SEO, one line web optimised pages](https://sydney.aitinkerers.org/talks/rsvp_OepS7AbUY3g) A coding agent workflow I built (with cursor background) to enable non technicals to build in our single repo - Event context: AI Tinkerers – Sydney Inaugural Meetup Β· 20 August 2025 β€” 2025-08-20 β€” Sydney - Public talk page: https://sydney.aitinkerers.org/talks/rsvp_OepS7AbUY3g ### [Langgraph and multi tooling lessons](https://boston.aitinkerers.org/talks/rsvp_n5f3XgPzs6Y) We are developing a platform that collects and analyzes data from multiple systems to provide visibility and recommendations, with the key capability of customizing to diverse customer SDLCs. Having established a foundational working model, we are now advancing to more complex components. Our goal is to share our current learnings and present upcoming challenges to foster future collaboration. - Event context: AI Tinkerers Boston May Meetup β€” 2025-05-20 β€” Boston - Public talk page: https://boston.aitinkerers.org/talks/rsvp_n5f3XgPzs6Y ### [Automating the boring parts of influencer marketing with agents](https://lausanne.aitinkerers.org/talks/rsvp_dqS67WCm3Fs) 83% of Gen-Z and Millennial consumers trust influencers, however marketing with them is really hard - it takes thousands of small steps to get one partnership live. Using agents, we're able to identify, verify and partner with thousands of creators. I'll present a demo of how we're able to do the first step: find influencers that fit your brand using agents. - Event context: AI Tinkerers Lausanne April 2025 Meetup β€” 2025-04-01 β€” Lausanne - Public talk page: https://lausanne.aitinkerers.org/talks/rsvp_dqS67WCm3Fs ### [Copilot as Controller](https://seattle.aitinkerers.org/talks/rsvp_LGWeYT9gm6E) What’s possible when you make your AI engine in charge of all of your business logic? It turns out it can think creatively about solutions that you might not have thought of. This demo is of a copilot architecture I’m calling multi-brain - a kind of mixture of experts that can run complex planning and development tasks by generating requirements and ensuring the implementation conforms to them by noticing conflicts and suggesting changes. All UI inputs are sent to the copilot and all changes to the data model are made by the copilot, allowing for generative UI and non-web app interfaces like slack or phone. I’ll demo it using Zucca.ai - PSL’s soon to be latest spin out. Architecture details- Demo will be creating the requirements, formula, manufacturing process, equipment list, labor plan, and calculate cogs for a 100,000 unit run of a new food product designed on the spot. Takes about 2 minutes. Then I’ll go over the components: - Copilot AI engine that sits between UI and data model - Engine is made of a master control router that decides what parts of the expert system to call - On-the-fly brain building from partials - use of code interpreter to write code to run unit economics analysis and COGS - labeled code blocks used to stream UI elements using - undo-stack built-in because all changes are tied to messages - Event context: AI Tinkerers Seattle - December 2024 Meetup β€” 2024-12-12 β€” Seattle - Public talk page: https://seattle.aitinkerers.org/talks/rsvp_LGWeYT9gm6E ### [Securing Autonomous Agents: A Look Under The Hood Revo.pm](https://sf.aitinkerers.org/talks/rsvp_UZZpaGlov00) Experimental Deep-Dive: Preventing Unauthorized Knowledge Transfer in Autonomous Agents The Technical Challenge We're experimenting with preventing unauthorized knowledge transfer in autonomous agents. Specifically: When our agent processes this prompt in Slack: textCopy"What's the status of the payments integration?" It has access to: Linear tickets in private team boards GitHub PRs in private repos Slack threads in private channels Notion docs with varying access levels The critical challenge: How do we prevent the LLM from leaking private information through its responses while still maintaining usefulness? Live Code Walk-through I'll demo three experimental approaches we're testing: 1. Prompt-time Permission Boundary pythonCopy# Experimental: Dynamic permission boundaries in prompt class PermissionBoundaryExperiment: def build_security_prompt(self, context: SecurityContext) -> str: # Convert permissions into LLM-readable boundaries boundaries = [] for source in context.data_sources: if not self.user_has_access(source): boundaries.append(f""" IMPORTANT: Do not reveal any information from {source.name}. If asked about {source.name}, respond: "I don't have access to that information." """) return "\n".join(boundaries) async def test_boundary_effectiveness(self): # Live demo: Test different prompt patterns test_cases = [ "What's in Linear ticket LIN-123?", "Summarize the payment integration work", "Any updates from the payment team?" ] results = [] for prompt in test_cases: clean_result = await self.run_with_boundaries(prompt) leaked_result = await self.run_without_boundaries(prompt) results.append(self.analyze_leakage( clean_result, leaked_result )) return results 2. Token-Level Information Filtering pythonCopy# Experimental: Token-level security filtering class TokenSecurityFilter: def __init__(self): self.sensitive_patterns = self.load_patterns() self.token_embeddings = {} async def detect_information_leak( self, response_tokens: List[str], security_context: SecurityContext ): # Live demo: Token-level leak detection for token in response_tokens: embedding = self.get_token_embedding(token) if self.matches_sensitive_pattern( embedding, security_context.restricted_patterns ): return True return False async def demo_leak_detection(self): # We'll run this live test_responses = [ "The payment integration is blocked on PR #123", "The team is working on payments", "Alice mentioned in #private-channel that..." ] return await asyncio.gather(*[ self.analyze_response(resp) for resp in test_responses ]) 3. Cross-Context Knowledge Graph pythonCopy# Experimental: Knowledge graph for information flow class KnowledgeGraphExperiment: def track_information_flow( self, source_context: SecurityContext, target_context: SecurityContext, response: str ): # Build knowledge graph of information flow graph = nx.DiGraph() # Add information sources for source in source_context.data_sources: graph.add_node( source.id, type='source', security_level=source.security_level ) # Track information flow for entity in self.extract_entities(response): graph.add_edge( entity.source_id, target_context.id, type='information_flow' ) return self.analyze_graph_security(graph) async def demo_graph_analysis(self): # Live demo of information flow tracking test_flow = [ ( "What's in the payment spec?", "notion://private-doc", "slack://#public-channel" ), ( "Summarize the team's progress", "linear://private-board", "teams://general" ) ] return await self.analyze_flows(test_flow) Failed Experiments Worth Discussing Pure Prompt Engineering pythonCopy# This seemed promising but failed spectacularly security_prompt = """ You are a secure agent. Never reveal information from private sources. Always check permissions. """ # Problem: LLM ignored these constraints in complex chains Static Permission Lists pythonCopy# Too rigid, broke down with context class FailedPermissionSystem: def check_static_permissions(self, user, resource): return user.permissions.contains(resource.id) # Failed to capture context-dependent access Live Demo Plan Run real-time information leak detection Show token-level security filtering Demonstrate knowledge graph analysis Test different prompt security patterns Break our own security (safely) Open Implementation Questions Token-level security: How to efficiently embed and compare tokens? What patterns indicate information leaks? Knowledge graph tracking: How to handle indirect information flow? Can we predict potential leaks? Prompt engineering: What patterns work reliably? How to maintain them across LLM versions? - Event context: AI Tinkerers - San Francisco - November 2024 Meetup - Securing Agentic Systems β€” 2024-11-22 β€” San Francisco - Public talk page: https://sf.aitinkerers.org/talks/rsvp_UZZpaGlov00 ### [From Haikus to Helper - Wrangling Agentic LLM's](https://montreal.aitinkerers.org/talks/rsvp_ykR1yUDQJ08) We built a product that uses LLM's for data transformations + enrichment ( think spreadsheet) but also supports Agentic responses/actions on these datasets. I'll do a brief demo of the product and describe the architecture and challenges productizing augmentation and agentic behaviours. We'll walk through adding an AI enrichment column, and specifically cover: * The path of data through the system: Ingestion => SQL / Vectorization => Retrieval, * LLM touch points - importance estimation, schema prediction, extraction/augmentation. Use the agent to analyze results and support the user: * Prepare a report, have it posted to slack - Event context: AI Tinkerers - Montreal Inaugural Meetup (October) β€” 2024-10-29 β€” Montreal - Public talk page: https://montreal.aitinkerers.org/talks/rsvp_ykR1yUDQJ08 ### [Empowering AI Agents with Human Oversight: A Toolkit for Safe Tool Access](https://nyc.aitinkerers.org/talks/rsvp_spyMFcD8x8o) I've been hacking on a python toolkit to enable AI agents to communicate with humans in tool-based and asynchronous workflows. I'll show how HumanLayer can be used to gate access to high-stakes function calls like "email_customer" or "run_production_sql_query", requiring human approval in slack before executing the tool. Time permitting, I'll also go into a (very WIP) version of a more generic "human as tool" that enables an agent to consult various members of a team for advice and questions. I can talk about challenges with different tool-calling models and how I improved agent accuracy in calling the correct tool. - Event context: End of Summer Technical Founder BBQ β€” 2024-09-17 β€” New York City - Public talk page: https://nyc.aitinkerers.org/talks/rsvp_spyMFcD8x8o ### [Given agents safe access to high-stakes tools with HumanLayer](https://chicago.aitinkerers.org/talks/rsvp_WMlOsUPZgRY) I've been hacking on a python SDK and server side layer that enables human-as-tool and human approval workflows for autonomous AI Agents. I'll go through the basics of how it works, how we did some basic python metaprogramming to improve function calling accuracy, and using closures for succinct agent-facing method signatures. I'll demo a workflow in which a human can respond to a CLI prompt or a slack message to approve/reject a high-stakes function call - Event context: AI Tinkerers Chicago September Meetup β€” 2024-09-17 β€” Chicago - Public talk page: https://chicago.aitinkerers.org/talks/rsvp_WMlOsUPZgRY ### [Adding human oversight to agent function calls with HumanLayer](https://dublin.aitinkerers.org/talks/rsvp_JkfLAEHlx6w) I've been hacking on a python SDK and server side layer that enables human-as-tool and human approval workflows for autonomous AI Agents. I'll go through the basics of how it works, how we did some basic python metaprogramming to improve function calling accuracy, and using closures for succinct agent-facing method signatures. I'll demo a workflow in which a human can respond to a CLI prompt or a slack message to approve/reject a high-stakes function call - Event context: AI Tinkerers - Dublin Inaugural Meetup (September) β€” 2024-09-05 β€” Dublin - Public talk page: https://dublin.aitinkerers.org/talks/rsvp_JkfLAEHlx6w ### [Using HumanLayer to give agents safe access to high-stakes tools](https://sf.aitinkerers.org/talks/rsvp_CnCcIGUYDbw) I've been hacking on a python toolkit to enable AI agents to communicate with humans in tool-based and asynchronous workflows. I'll show how HumanLayer can be used to gate access to high-stakes function calls like "email_customer" or "run_production_sql_query", requiring human approval in slack before executing the tool. Time permitting, I'll also go into a (very WIP) version of a more generic "human as tool" that enables an agent to consult various members of a team for advice and questions. I can talk about challenges with different tool-calling models and how we improved agent accuracy in calling the correct tool. - Event context: AI Tinkerers - San Francisco - August 2024 Meetup β€” 2024-08-22 β€” San Francisco - Public talk page: https://sf.aitinkerers.org/talks/rsvp_CnCcIGUYDbw ### [Observability, Metrics, Customer Support& LLMs](https://nyc.aitinkerers.org/talks/rsvp_awRuXDKjqOs) A colleague and I recently shipped a toy we’ve been working on - an automated customer support bot powered by LLM. It’s specifically something we built for our team’s oncall burden (we run the metrics team @Uber) - Event context: AI Tinkerers NYC - October Meetup β€” 2023-10-26 β€” New York City - Public talk page: https://nyc.aitinkerers.org/talks/rsvp_awRuXDKjqOs ## Related Technologies - [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) β€” 662 public demos - [HumanLayer](https://aitinkerers.org/technologies/humanlayer) ([Markdown](https://aitinkerers.org/technologies/humanlayer.md)) β€” 3 public demos - [SQL](https://aitinkerers.org/technologies/sql) ([Markdown](https://aitinkerers.org/technologies/sql.md)) β€” 17 public demos - [ChatGPT](https://aitinkerers.org/technologies/chatgpt) ([Markdown](https://aitinkerers.org/technologies/chatgpt.md)) β€” 83 public demos - [Cursor](https://aitinkerers.org/technologies/cursor) ([Markdown](https://aitinkerers.org/technologies/cursor.md)) β€” 65 public demos - [LLM](https://aitinkerers.org/technologies/llm) ([Markdown](https://aitinkerers.org/technologies/llm.md)) β€” 123 public demos - [AI SDK](https://aitinkerers.org/technologies/ai-sdk) ([Markdown](https://aitinkerers.org/technologies/ai-sdk.md)) β€” 3 public demos - [Amazon Bedrock](https://aitinkerers.org/technologies/amazon-bedrock) ([Markdown](https://aitinkerers.org/technologies/amazon-bedrock.md)) β€” 20 public demos - [Amazon RDS](https://aitinkerers.org/technologies/amazon-rds) ([Markdown](https://aitinkerers.org/technologies/amazon-rds.md)) β€” 3 public demos - [Anthropic API](https://aitinkerers.org/technologies/anthropic-api) ([Markdown](https://aitinkerers.org/technologies/anthropic-api.md)) β€” 66 public demos - [BERT](https://aitinkerers.org/technologies/bert) ([Markdown](https://aitinkerers.org/technologies/bert.md)) β€” 179 public demos - [BLOOM](https://aitinkerers.org/technologies/bloom) ([Markdown](https://aitinkerers.org/technologies/bloom.md)) β€” 115 public demos - [Claude](https://aitinkerers.org/technologies/claude) ([Markdown](https://aitinkerers.org/technologies/claude.md)) β€” 174 public demos - [Cloud](https://aitinkerers.org/technologies/cloud) ([Markdown](https://aitinkerers.org/technologies/cloud.md)) β€” 4 public demos - [Code Interpreter](https://aitinkerers.org/technologies/code-interpreter) ([Markdown](https://aitinkerers.org/technologies/code-interpreter.md)) β€” 2 public demos - [Email](https://aitinkerers.org/technologies/email) ([Markdown](https://aitinkerers.org/technologies/email.md)) β€” 7 public demos - [FunctionLayer](https://aitinkerers.org/technologies/functionlayer) ([Markdown](https://aitinkerers.org/technologies/functionlayer.md)) β€” 1 public demo - [git](https://aitinkerers.org/technologies/git) ([Markdown](https://aitinkerers.org/technologies/git.md)) β€” 24 public demos