All Documents
66 documents available
LlmGuard Framework - Complete Implementation Buildout
Specifies a complete 24-week buildout plan for an Elixir-based AI firewall and guardrails framework with multi-layer threat detection.
Agent Security and Interoperability
Defines security and interoperability standards for enterprise agentic AI deployments, covering authentication, authorization, and agent interaction protocols.
Guardrails, Safety & Content Filtering
Implements a layered guardrail system with input validation, output validation, and content filtering to protect LLM applications from prompt injection, jailbreaks, and data leaks.
TODO (ConcreteSky)
Maps the full feature backlog for a Bluesky community groups PHP package, from done items to planned panels and parity targets.
AI Workforce Playbook
Shares real-world lessons from building an AI agent team with OpenClaw, including scoring, autonomy phases, pipeline design, and failures.
AI Red Teaming Workshop - Discovery & Attack Demonstration Guide
Guides a hands-on workshop where participants simulate four LLM attack types using Microsoft's official red teaming labs.
Phase 1 Test Implementation - Review Guide
Guides code review of 9 test files (250+ cases) for patient safety modules, with run instructions and expected outcomes.
NOTES
Guides building a public-facing water quality app with map integration, data processing, and stakeholder-specific presentation.
Decision Trees
Explains decision tree concepts: entropy, information gain, ID3 algorithm, overfitting, and handling continuous or many-valued attributes.
Private Advertising Technology Working Group / Community Group Minutes - 2025-06 Meeting
Records discussion and decisions from a W3C working group meeting on global safety limits and per-site budgeting for Private Advertising Technology.
agent-CLAUDE
Defines the identity, workspace layout, sync rules, and behavioral constraints for a long-lived Claude Code session acting as a company knowledge agent.
WEB:OS — The Web Content Operating System
Defines a branded boot sequence, workspace structure, execution modes, and 30+ slash commands for a web content operating system.
Ads Agent – RinkLink
Defines an autonomous paid ads agent that creates, executes, and optimizes Meta, Google Ads, and X campaigns for a hockey scheduling product.
LiftReel Terms of Service (including Software License/EULA)
Defines the legal contract for a fitness social app, including a software license, user obligations, and dispute resolution terms.
Midjargon Package Implementation Plan
Lays out a multi-priority plan to fix Python version compatibility, Pydantic models, parser, CLI, and test suite for a Midjourney prompt parsing package.
Agent Design Fundamentals
Defines core agent components, control loop pseudocode, autonomy levels, memory strategies, safety considerations, and design anti-patterns for building AI agents.
Growstuff Terms of Service
Sets out the legal terms for using the Growstuff gardening service, covering accounts, content, conduct, and liability.
英文隱私權條款範本文件
Provides a standard English privacy policy template with placeholders for website name and contact details.
GangGPT - AI-Powered GTA V Multiplayer Server
Describes an AI-powered GTA V multiplayer server built on RAGE:MP with Azure OpenAI, covering features, architecture, and deployment.
Risk Assessment Matrix
Documents 12 risks with likelihood, impact, score, level, and mitigation for a child-facing AI app's data protection impact assessment.
Twitter/X Launch Thread
Provides a 6-tweet launch thread template with timing, replies, amplification strategy, and success metrics for a GitHub project.
📰 AI News Daily — 09 Dec 2025
Summarises 09 Dec 2025 AI news across tools, LLMs, research, policy, and discussions for a daily briefing.
Implementing AI-Safety in a LLM-System Architecture
Explains hallucination risks in LLMs and proposes an AI Safety Module with content filters for business customer-service use.
DeepSeek R1: Case Study in Failed Extrinsic Alignment
Compiles public security research and independent findings to argue that extrinsic alignment methods are insufficient for AI safety, using DeepSeek R1 as a case study.