LLM Mart Basic

Use when compiling for ARM, AArch64, RISC-V, or MIPS from x86-64 with a GCC cross toolchain: triplets, sysroots, pkg-config, CMake toolchain files, QEMU. Not for Zig: use zig-cross.

Use when one reviewer is not enough because failure modes are heterogeneous, or a claim needs cross-lens pressure before it ships. Not for collapsing a decision field: use converge.

Use when a long agentic project needs each cycle to end in a learning memo and a keep/iterate/restart decision. Not for single-pass builds.

Use when asked for a sequence diagram of cryptographic protocol semantics from code, prose, RFCs, papers, ProVerif, or Tamarin, or for code/spec divergence. Not for architecture: use diagramming-code.

Use when writing CUDA kernels, managing the thread, block, and grid hierarchy, tiling shared memory, using streams, setting nvcc flags, or using Thrust. Not for kernel debugging: use cuda-debugging.

Use when debugging CUDA with cuda-gdb or Compute Sanitizer, reading GPU core dumps, using device printf, or triaging error codes 700, 701, 702, and 719. Not for performance: use cuda-profiling.

Use when profiling CUDA with Nsight Systems or Nsight Compute, reading roofline and occupancy metrics, or annotating phases with NVTX. Not for correctness: use cuda-debugging.

Use when Survey and Job Culture Index profiles need analysis for stress, burnout, disengagement, or flight-risk signals. Not for clinical diagnosis: use a qualified clinician.

Use when two colleagues' working friction needs trait-based explanation, accommodations, process changes, and escalation boundaries, or for manager-report friction. Not for performance adjudication.

Use when a Culture Index profile needs comparison with role requirements, team composition, and manager profile for hiring. Not for transcript prediction: use culture-interview-profile-prediction.

Use when asked to predict Culture Index traits from an interview transcript before a survey exists, including sparse or contradictory evidence. Not for interpreting completed survey results.

Use when a manager needs profile-specific communication, one-on-one, motivation, and energy guidance for a direct report. Not for general pair or team compatibility analysis.

Use when a signed new hire's Culture Index profile and team profiles need a first-90-days plan. Not for manager coaching: use culture-manager-coaching.

Use when implementing pool/slab/arena allocators, tuning jemalloc/mimalloc/tcmalloc, writing a Rust GlobalAlloc, or benchmarking allocator performance and fragmentation.

Use when customer feedback, NPS, churn, email feedback, call transcripts, or voice-of-the-customer analysis needs a report over a time window.

Use when the user asks to cut, trigger, or start a release candidate for a release branch. Not for full releases, hotfixes, or non-release-candidate workflow dispatches.

Use when the caller supplies a falsifiable out-of-happy-path invariant and a finite budget. Restores via bounded patches or reverts. Not for normal feature delivery or universal retries.

Use when a build, QA pass, demo, user complaint, or abandoned attempt leaves the next pass needing lessons rather than code. Not for changelog extraction. Not for session handoff: use handoff.

Use when a human-curated dbt model index must guide BigQuery SQL for a warehouse question. Not for discovering undocumented models or executing warehouse changes.

Use when a user asks to tighten verbose-but-correct prose without a full rewrite. Cuts in place to load-bearing density, preserving every load-bearing claim.

Fourteen posts of being wrong in production, compressed to checkboxes

Healthy nodes, a quiet network, 300 restarts in three days, and a latency budget measured in milliseconds

Discovery worked. Ping worked. Every TCP connection timed out, and later the tunnel only worked when someone had a terminal open.

Every VM came back. The cluster did not. Declarative systems converge on config, and the datapath isn't config.

A surprising share of AI-in-the-terminal failures aren't the AI. They're zsh, and a version of bash from 2006.

A Claude Code plugin turns standalone project configuration into a namespaced, installable extension that teams and communities can update as one unit.

None of the safety came from the model. It came from six boring habits.

Skills package instructions and references. Subagents run work in a separate context and return results. They solve different problems and can be composed deliberately.

Six hours in, one step left, everything green, and the incident that didn't happen

CLAUDE.md carries persistent project context. Skills load reusable procedures when relevant. Separating stable facts from task-specific workflows keeps both easier to maintain.

Twenty minutes recovering secrets that never existed, and the one sentence from a human that ended it

An API request routing a model's tool call through an approval gate to a remote MCP server

31 config keys, two audits, and why the first one was wrong in both directions

The official MCP Registry stores standardized server metadata rather than package code. Publishers verify a namespace, describe installation or remote access, and submit immutable versions.

Everyone looks at the Dockerfile. The file that actually leaked the key was the project file.

Remote MCP authorization uses established OAuth standards, but secure integration still requires issuer validation, least-privilege scopes, protected token handling, and server-side enforcement.

"Copy it over and switch the reference" is two steps, and the outage lives in the one nobody checks

stdio fits local processes and prototypes. Streamable HTTP fits hosted services and shared integrations. The right choice follows where the capability runs and who must reach it.

The most important rule wasn't about what I could change. It was about what I was allowed to display.

Tools perform operations, resources expose readable context, and prompts provide reusable templates. Choosing the correct primitive makes an MCP server easier to understand and govern.

This is the lane for changes with no behavior to test-drive — prose edits, a version bump on an

Use this after a pull request has merged but your local checkout is still on the topic branch.

Prints the public command catalog straight from `COMMANDS.md` — the plugin's own single source

This is the single entry point for turning staged work into a commit — nothing in codeArbiter

The protocol for a rule conflict — not a skill route, an orchestrator-level halt. When two sources

An optional, on-demand drift audit for the bypass case: a merge, a direct push, or a manual edit

This is the populator for a project that already has code to read. Instead of interviewing you about

`/ca:debug` investigates an unexplained symptom and returns a cited diagnosis or a bounded

This is the populator for a project that has no code yet to read. Rather than guessing at

Proves the install is actually enforcing, rather than just present. codeArbiter's worst failure

This is the standard entry point for new work with a human in the loop at every step. A short

This is the entry point for an authorized repair of a confirmed code defect. Give the observed

This is how a repository opts into codeArbiter for the first time. It writes the root-level state

A bare-numbers governance glance — three metrics, each with a trend arrow against the prior

The sanctioned, logged escape hatch. A routine gate — a lint rule, a style check, a non-security

Explicit PR entry and a direct request to open a PR use the same branch-finishing owner.

A zero-onboarding, read-only dry-run of the reviewer fleet against whatever is currently

This is a Feature Forge preview command — the after-each-turn service ships **off** by default and

Compares architectural records with the scaffold and prior decisions using SMARTS.

This is the lane for moving or reshaping code without changing what it does — a rename, an extract,

Offline-first Python AI agent that runs a tiny research business: quotes each job against its own costs, collects via Stripe, fulfils with NVIDIA Nemotron, pays…

DSH 插件 · 注入式优化器 0.8（主线）：你照常说话，它在你发送后,AI接收前把"这一轮到底要什么"理清楚，再把这份理解交给工作 AI(上下文注入) —— 原话不改写，条条带逐字依据。含控制界面（档位/权限/模型/上下文/只读工具）、拦截浮层（思维层+产出层）与真实 token 用量。可明显提升大多数模型的发挥稳…

Reliable AI Skill + MCP toolkit for agent-driven desktop CAD automation.

Curated real-world use cases for Hermes Agent — the self-improving AI agent from Nous Research. Backed by primary sources.

AI Agent 教学仓库 | 系统化 LangChain、RAG、LangGraph、MCP 全栈实战代码 | 万字博客详解 | 开源可运行示例 | 从零构建智能体

从 0 复刻 WorkBuddy-style 桌面 AI 助手 Harness：24 章 Python 教程，覆盖 Agent Loop、工具调用、记忆系统、Sidecar、沙盒审计、DeepSeek/OpenAI 评测轨迹

MCP server and CLI tools for web search and crawling, built on SearXNG and Crawl4AI

Delta MCP is a free app that sits between your AI apps and their MCP servers: one program per task instead of one tool call per step, up to 24.1× fewer tokens i…

Ava turns any Android 5+ device into a voice-first Home Assistant kiosk. Native C++ under the hood, so a 10-year-old tablet still listens, talks, and runs the h…

Roblox Studio macOS Window Capture Fix 2026: Real Screenshot Tool Instead of Magenta Playtest Glitch

LabTether