You are working on a project that uses Squad, an AI team framework. When picking up issues autonomously, follow these guidelines. Before starting work on any issue: Before starting work, check your capability profile in .squad/team.md under the Coding Agent → Capabilities section. Use the squad branch convention: Example: squad/42-fix-login-validation When opening a PR: - Reference the issue: Closes #{issue-number} - If the issue had a squad:{member} label, mention the member: Working as {member} ({role}) - If this is a…
You are working on a project that uses Squad, an AI team framework. When picking up issues autonomously or using Squad, follow these guidelines. IMPORTANT — run this check only when the session is expected to use Squad Coordinator behavior, before doing any Squad work: Look for the exact token SQUADCOORDINATORCANARYa8f3 in your agentinstructions. This token is placed at the end of squad.agent.md. If it is missing, the coordinator file is missing or truncated. Do not run this check for…
Bernstein is the open-source governance layer for AI agents (Python 3.12+), one git worktree per task. There is no model in the coordination loop - the orchestrator is deterministic Python, NOT an LLM - so the same plan replays to a byte-identical task graph. NEVER run uv run pytest tests/ - it leaks 100GB+ RAM. Always use scripts/run_tests.py.
This repository uses Aegis as method-pack workflow discipline. Before risky implementation, prefer the smallest relevant Aegis skill from .github/skills/ or the linked Aegis skill directories. Keep detailed workflow logic in the skill body rather than duplicating it here. Key repository boundaries: - treat Aegis as Aegis Method Pack (runtime-ready), not a runtime core - keep user instructions and target-project rules above Aegis guidance - do not claim authoritative GateDecision or final completion authority - require fresh verification evidence before completion…
Always reference these instructions first and fallback to search or bash commands only when you encounter unexpected information that does not match the info here. Jupyter MCP Server is a Python-based Model Context Protocol (MCP) server implementation that enables real-time interaction with Jupyter Notebooks. The project uses a modern Python build system with hatch, and includes comprehensive testing, linting, and documentation. NETWORK TIMEOUT WARNING: pip install commands may fail with SSL certificate errors or read timeouts when connecting to PyPI.…
NEVER commit, push, or create PRs without explicit user instruction. NEVER auto-fix issues found during code review. Report findings only. NEVER present code as complete without running Invoke-CodeFormatterOnFiles and SpellCheck.ps1 first. ALWAYS report findings and wait for user decision before modifying the repository. ALWAYS commit only the specific files the user requests. No extra files. ALWAYS use named parameters instead of positional parameters. ALWAYS ensure .ps1 files have UTF-8 BOM encoding. When counting code patterns (function calls, occurrences), ALWAYS exclude…
Griffel is a CSS-in-JS library with near-zero runtime, SSR support, and ahead-of-time compilation. It's organized as a monorepo using Nx with 16+ packages and applications including React bindings, build-time transforms, and DevTools. Always reference these instructions first and fallback to search or bash commands only when you encounter unexpected information that does not match the info here. CRITICAL: NEVER CANCEL long-running commands. Set timeouts appropriately. When using Nx affected commands in a repository that may not have a main branch…
Prompty is being rebuilt from the ground up. The type system that originated in Prompty's TypeSpec specification is now owned in this repo and emitted through Typra. Prompty v2 uses the generated Prompty model packages as the canonical, lower-level type layer. Core principle: Prompty is a markdown file format (.prompty) for LLM prompts. The frontmatter uses Prompty's Typra-generated prompt model. The markdown body becomes instructions. The runtime loads, renders, parses, and executes prompts. Only the prompt runtime subset — not…
Review changes as an Agent Skills expert and Xcode build optimization reviewer. Keep this repository focused on: Do not turn the repository into general iOS architecture guidance or product documentation. When reviewing changes, prioritize:
This repository follows a structured workflow based on two primary branches: All work must follow the branching, versioning, and changelog rules defined below. Unless explicitly instructed otherwise: Examples: - feature/conversation-metadata - feature/improve-chat-history Examples: - release/1.2.3 - release/2.4.2 When preparing a release branch: - Update version references where applicable - Update CHANGELOG.md - Ensure the repository reflects a releasable state - Do NOT introduce new feature work - GitHub Release titles MUST be exactly the tag name (for example, v2.8.0). Never…
Microsoft.Diagnostics.Runtime (ClrMD) is a .NET library for inspecting crash dumps and live processes. It provides programmatic access to CLR internals (heap, types, threads, GC roots, exceptions) similar to SOS debugging extensions. Important: ClrMD loads architecture-specific native components (DAC), so the test host must match the dump architecture. Use --arch x86 to test 32-bit dumps and --arch x64 for 64-bit. The Test.cmd/test.sh scripts handle this automatically. Dump generation: Tests generate crash dumps lazily on first run — no manual pre-build step…
This project is a monorepo implementing the MCP (Model Context Protocol) server that lets users interact with their MongoDB clusters and MongoDB Atlas accounts. It is built using TypeScript, Node.js, pnpm workspaces, and the official @modelcontextprotocol v2 SDK packages (client, server, core, node). The server ships as the v3 API: a set of scoped @mongodb-js/mcp-* packages. The mongodb-mcp-server package is the binary CLI distribution (npx mongodb-mcp-server / MCPB) — library embedding uses the scoped packages. This project uses pnpm workspaces…
This project is the Microsoft JDBC Driver for SQL Server, enabling Java applications to interact with SQL Server and Azure SQL databases. It implements the JDBC 4.2/4.3 specification, communicating over the TDS (Tabular Data Stream) protocol. It supports features like connection pooling, Always Encrypted, Azure AD authentication, bulk copy, configurable retry logic, and idle connection resiliency. The project builds from a single Maven project (pom.xml) with multiple JRE profiles (jre8, jre11, jre17, jre21, jre25, jre26). Each profile compiles against a…
This is ONNX Runtime GenAI, a high-performance inference library for generative AI models. The codebase implements the complete generative AI loop including preprocessing, ONNX Runtime inference, logits processing, search/sampling, and KV cache management. The State class hierarchy in src/models/model.h handles device-specific execution, while the Engine class in src/engine/ manages request batching and scheduling. WinML builds require explicit SDK version specification: WinML integration downloads Microsoft.Windows.AI.MachineLearning via NuGet and copies headers/libs to a local ort/ directory. Each hardware backend implements DeviceInterface (defined…
This repository's canonical agent context lives in AGENTS.md at the repository root. Read it first. It describes the repository layout, the script conventions (docs/conventions.md), the one-command verify loop (scripts/verify.sh), and the pull-request conventions (.github/instructions/telemetry.instructions.md). These are sample scripts for educating Intune administrators; they are run in non-production / test environments only. Keep changes focused, idempotent, and consistent with the existing ## banner plus Microsoft copyright / MIT license header style. Run ./scripts/verify.sh before opening a pull request.
The following instructions are only to be applied when performing a code review. Review the current pull request diff for security, code quality/readability, performance, test coverage, and user impact. Only…
This is a .NET based repository that contains the MSTest testing framework and Microsoft.Testing.Platform (aka MTP) testing platform. Please follow these guidelines when contributing: The codebase ships several distinct (but related) products. Knowing which product a change belongs to is essential because they have different conventions, target frameworks, and public API surfaces: Solution files: TestFx.slnx is the full solution; MSTest.slnf, Microsoft.Testing.Platform.slnf, and NonWindowsTests.slnf are filtered views. Always use the repo-local toolchain via the build scripts — they restore the pinned…
MicrosoftCopilot instructions
