LLM Mart Basic

Monitor the CI pipeline for the current branch via a background Monitor script (GitHub or GitLab), reacting to pass, fail, and manual-gate states. Use when the user says 'watch CI', 'monitor the pipeline', 'is CI green', or after pushing a branch or creating a PR/MR.

Runs a structured production-incident investigation that forces evidence-first hypothesis ranking before any code change. Use when given an error message, Sentry alert, failing log, or an 'investigate <X>' request.

Creates or updates a diagram, picking mermaid vs drawio per rules/diagrams.md, writing the source file, and previewing via MCP. Use when the user says 'diagram' or '/diagram', or asks for a flowchart, architecture, sequence, or state diagram.

Drives a fleet of MRs/PRs to done with a manager loop plus the built-in /goal command, delegating all edit, review, rebase, and conflict work to worktree-isolated domain-expert subagents. Use when the user says 'drive fleet' or 'drive the fleet', has 2+ independent lanes to drive

Investigates and fixes a GitHub issue. Use when given an issue number or URL, or when the user says 'fix issue'.

Runs a grilling session that challenges a plan against the existing domain model, sharpens terminology, and updates the CONTEXT.md glossary inline as decisions are made. Use when the user wants to stress-test a plan against their project's language and documented decisions.

Compacts the current conversation into a handoff document another agent can pick up. Use when the user says 'handoff', 'hand off', or wants to continue this work in a fresh session.

Finds deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI

Read and write Jira work items through the acli CLI. Use when the user mentions a Jira ticket, issue, story, bug, or epic, drops a Jira key like SER-123, or pastes an atlassian.net/browse URL.

Create a merge request or pull request from the current branch: verify quality gates, enforce conventional commits, validate the title, and create it without a confirmation step. Use when the user says 'create MR', 'create PR', 'open a merge request', or 'raise a PR'.

Builds a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.

Reviews code exclusively for over-engineering and lists what to delete: reinvented standard library, unneeded dependencies, speculative abstractions, dead flexibility. Use when the user says 'review for over-engineering', 'is this over-engineered', or invokes /prune. Complements

Starts Phase 1 (Research) for a topic: reads every relevant file, optionally runs a panel of Explore teammates, and saves a research artifact to .claude/state/research/. Use when the user says 'research <topic>' or '/research', or before planning work in unfamiliar code. Research

Reviews a pull request with structured severity-based feedback. Use when asked to review a PR, asked for a code review, or given a PR number/URL.

Fetches and digests Sentry issue data (summary, tags, stack trace, breadcrumbs, latest event) by short ID, numeric issue ID, or sentry.io URL, for any Sentry org the local token can access. Use when the user mentions a Sentry issue or short ID (e.g. MY-PROJECT-4X2), pastes a sent

Runs pre-launch validation and the release workflow. Use when the user says 'ship', 'release', 'deploy', or 'ready to merge'.

Summarizes the current session's work into a diary entry at .claude/state/sessions/ and runs worktree auto-cleanup. Use when the user says 'summarize' or '/summarize', or when closing out a completed work session.

Breaks a plan, spec, or PRD into independently-grabbable issues on the project issue tracker using tracer-bullet vertical slices. Use when the user wants to convert a plan into issues, create implementation tickets, or break down work into issues.

Runs the comprehensive quality gate before declaring work done: discovers the checks CI actually runs, executes them in order, then reviews git status and the session diff. Use when the user says 'verify done' or '/verify-done', or before pushing any branch.

Plans a huge chunk of work - more than one agent session can hold - as a shared map of decision tickets in a local file, and resolves them one at a time until the way to the destination is clear. Use when the user invokes /wayfinder on an effort too big for a single session.

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.

`crabbox attach` follows the recorded events of an active coordinator run and

`crabbox azure` groups Azure provider setup commands. It currently has a single

`crabbox bench` records and reports local benchmark timing observations. It is a

`crabbox cache` inspects, purges, or warms package and build caches on a

`crabbox capsule` captures, replays, and tracks lightweight failure capsules.

Save the state of a lease, then restore it onto another box or fork it into a

`crabbox claims list` prints the lease claims stored on the current machine. It

`crabbox cleanup` sweeps direct-provider machines and local provider state that

`crabbox code` bridges a Linux lease's `code-server` workspace into the

`crabbox config` inspects and updates user configuration. It has three

`crabbox connect` resolves a lease and opens an interactive SSH session to it.

`crabbox cp` copies files or directories between the host and a Crabbox-owned

`crabbox desktop` drives a visible desktop session on a lease that was warmed

`crabbox doctor` runs a preflight before you commit to a long workflow. It is

`crabbox egress` gives a lease mediated outbound network: a lease-local browser

`crabbox events` prints the broker's event log for a recorded run.

`crabbox heartbeat` refreshes the idle deadline for one owned lease and prints

`crabbox history` lists recorded remote command runs from the broker. Each run is

`crabbox image` holds the trusted-operator controls for provider base images:

`crabbox init` onboards the current repository: it writes the minimal config

AI agent orchestration kit for Windows, Linux/MacOS with Codex skills, hooks, routing rules and profiles for Claude, OpenCode, Cursor, Gemini and Windsurf.

Skills & reviewer agents for AI-first climate science — built and used by a PhD atmospheric scientist

出行路书工作流 skill：联网实查 + 多源交叉验证，产出可核验、能执行的旅行攻略。覆盖吃住行游拍避全维度，附美食情报卡、基准骨架与校验工具，支持一键部署在线版。

Self-hosted AI code reviewer with indexed PR reviews, walkthroughs, vulnerability scanning, dependency graphs, custom rules, and a learning loop.

🧬 Extend EvoScientist with Installable Skill & Knowledge Packs

Give the agent a machine. Just not yours. Each AI coding agent gets its own isolated machine with root, Docker, and systemd - active defense detects and stops t…