LLM Mart Basic

Apply whenever writing or updating a project-facing document — root CLAUDE.md, `.hedgehog/core-design.md`, specs, READMEs, or similar. Ensures the file reads as a clean, current, as-is snapshot rather than a log of edits or decisions. Applies on first generation and on every late

Push the LLM to reconsider, refine, and improve its recent output. Use when user asks for deeper critique or mentions a known deeper critique method, e.g. socratic, first principles, pre-mortem, red team.

Facilitate a brainstorming session using diverse creative techniques. Use when the user says 'help me brainstorm' or 'help me ideate'.

Decision-grade research, three ways: draft a deep-research prompt for the user to run in their own tool (ChatGPT, Gemini, Grok, Perplexity, …), process a finished research report — file it, distill a succinct cited summary with metadata that downstream skills consume without repr

Pressure-test an idea through persona-driven interrogation until it hardens, proves out, or dies cheaply. Use when the user says 'forge an idea', 'pressure-test this idea', 'stress-test my thinking', or 'harden this idea'.

Create, update, or validate a PRD. Use when the user wants help producing, editing, or validating a PRD.

Working Backwards PRFAQ challenge that stress-tests a product concept customer-first. Use when the user requests to 'create a PRFAQ', 'work backwards', or 'run the PRFAQ challenge'.

Create, update, or validate a product brief. Use when the user wants help producing, editing, or validating a brief.

Plan UX patterns and design specifications. Use when the user says "lets create UX design" or "create UX specifications" or "help me plan the UX"

Maintainer-only. Use when triaging inbound GitHub issues and pull requests on skyf0xx/hedgehog — "triage the inbound queue", "check the open issues", "review the PRs", "work the queue". Reads every item read-only, judges it for security and for merit, then fixes and closes or com

Use once per invocation, at the start of a new Hedgehog project, to land the workspace for whichever core `planner` selected at Phase 0 — the first real workspace this project gets, since `init` with no explicit core flag lands the shared agents/skills/build-graph payload and lea

Use for planning intake (core selection, then scope boundary + domain vocabulary or Chain Method brief, depending on core) at the start of a project, and for re-entry when new scope enters play on a project already built or mid-build — including after a build has reached its Stop

Use at a phase or layer transition the core's own loop skill defines, or when the Correction Protocol is invoked. Also use when the user asks for a review, audit, or "look over this". Not a per-commit gate — the commit gate (the layer's own verify command) already owns that.

Use once a core's build is complete (every task in the build graph `complete`) and the user is offered a fresh-context session to iterate. Takes post-build tweak requests one at a time from a clean context, and — separately — reviews accumulated build friction and asks the user d

Imported from skyf0xx/hedgehog/vendor-skills/BMAD/bmm-skills/plan/bmad-prfaq/agents/artifact-analyzer.md.

Imported from skyf0xx/hedgehog/vendor-skills/BMAD/bmm-skills/plan/bmad-prfaq/agents/web-researcher.md.

Review AI-generated or human-written code changes with fallow's graph-grounded review brief. Subtracts deterministic concerns (unused code, complexity, duplication, styling) from the loop, ranks what to look at by blast radius and risk, and surfaces the few consequential structur

Add local office-document-to-Markdown conversion to NanoClaw agent containers with the pinned Firecrawl AnyDoc CLI. Use when agents need to read attached Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV, or text-based PDF files without uploading them to a hosted parser.

Add Atomic Chat MCP server so the container agent can call local models served by the Atomic Chat desktop app via its OpenAI-compatible API.

Add clidash — a zero-dependency, read-only web dashboard that derives its tabs and tables at runtime from any CLI that lists resources as JSON. Ships pre-wired for NanoClaw's ncl CLI (agent groups, sessions, channels, users, roles), plus message-activity charts, a log tail, and a

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.

Record the present state by mode — decision (ADR, RFC, rule), code (spec, doc, guide, scenario), or research (a ready report or one external material); a gate picks the document type.

First-time Archcore setup — wire host configs, measure the authored context, compose the full first-day seed in one preview, and create it on one confirm; import converts CLAUDE.md, AGENTS.md, rule files, ADRs, and docs into native documents; refresh adds new facts or drills into one domain.

Plan a feature or initiative through a computed route — a small fix exits with no documents, one capability gets a spec and a plan, a large initiative gets an umbrella PRD with one spec per capability; start with sdd, sources (market research), iso (regulated work), or research (technical investigation) to run that path directly.

Review branch changes against Archcore docs, or report project health; drift runs staleness detection, deep a full documentation audit, closeout closes a finished feature, experience captures a repeated pattern.

Verify that citations actually exist and that the claims they support are faithful to the cited source. Runs deterministic existence checks (Crossref / OpenAlex / Semantic Scholar / arXiv) plus a claim-faithfulness pass via the alterlab-citation-verifier skill.

Run a systematic, reproducible literature review on a topic and return an APA 7.0 annotated bibliography with a documented search strategy. Invokes the alterlab-deep-research pipeline in lit-review mode.

Run a full multi-perspective peer review of a manuscript, simulating an Editor-in-Chief plus three peer reviewers and a Devil's Advocate, and produce a structured editorial decision and revision roadmap. Invokes the alterlab-paper-reviewer skill.

Orchestrate the end-to-end academic research-to-publication workflow (research, write, integrity check, review, revise, re-review, finalize) with mandatory integrity gates and two-stage peer review. Invokes the alterlab-research-pipeline orchestrator.

Audit infra security: secrets, deps, CI/CD, webhooks, AI/skill files

Audit Solana program code for exploitable bugs and write a findings report

Compare per-instruction CU with the stored baseline to catch regressions

Build the web client (Next.js, Vite, React) and check env, types and bundle

Build Solana programs (Anchor, Pinocchio, native), incl. verifiable builds

Build the Unity project in batchmode for WebGL, desktop, Android or PSG1

Turn a solana-ai-kit fork into a project: set up CLAUDE.md, remove kit files

Un-ignore and commit the kit config dir, instruction file, .mcp.json and .gitmodules

Replay a user's failing transaction on forked state and map the error to source

Deploy a program to devnet, or to mainnet after the user's explicit go-ahead

Review the branch diff for Solana security issues, CU waste and AI slop

Read-only check of toolchain and kit config, with one fix-it command per failure

Your car as a chat-room agent: Raspberry Pi 5 + dashcam + local AI. CodeWatch's sibling for the garage.

The enterprise gateway for autonomous agents. Identity management, per-channel isolation, credential vault, per-session tamper-evident audit log.

AI agent for Home Assistant — talk to your home, create automations in plain language, analyze cameras with face recognition, and get proactive anomaly alerts.…

A desktop app to prototype agent ideas, inspect every harness step, replay failures, and evaluate performance, all in one place. Local-first, cloud-ready for ma…

Local-first AI memory you can see, edit, and override — portable across Claude Code, Codex, Cursor, Windsurf, and other MCP coding tools.

Web & desktop (Electron) client for the OCTO open workplace — one React + TypeScript codebase shipping browser and PC surfaces, with first-class AI agent UX.

同花顺官方 A股金融数据服务，提供股票实时行情、历史行情、财务报表、指数、板块、涨停等数据，适用于 AI Agent、量化研究和应用开发，支持 API、MCP、CLI 和 Python。Official Tonghuashun (HiThink) A-share financial data service provi…

Open-source desktop SQL workspace for PostgreSQL, MySQL/MariaDB, SQLite and 15+ more databases like DuckDB, ClickHouse, Redis and Firestore. Built-in MCP server…

🤖 Local-assist BOSS Zhipin CLI for AI agents — search, welfare filtering, shortlist, JSON-envelope output; low-risk & compliant by default.

The local-first sidebar AI agent for ComfyUI — runs on your own Claude OR ChatGPT subscription (no API keys, no extra LLM costs). Drives your live graph: edits,…

Persistent memory for Claude Code — identity, context, and continuity across sessions

🔬 A curated collection of 23,000+ agent skills for empirical research across 8 social science disciplines. | 精选 23,000+ AI Agent 技能库，覆盖8大社会科学学科的实证研究。CoPaper.AI…

A curated list of plugins, skills, MCP servers, patch/profile layers, orchestrators & UIs for DeepSeek Harness (DSH). Visualization · PPT · Coding · Agents · Lo…

World Memory Protocol.

Run your own organization of agents.

The most RAM efficient harness

AI-native, free, open-source alternative to Jira, Trello, ClickUp & Monday. Built for Scrum teams where humans and AI agents collaborate as equals — on the same…

Free open-source desktop app that scrapes JAV metadata and generates NFO + cover art for Jellyfin, Emby & Kodi. No Docker, no CLI — one-click install on Windows…

A cross-platform desktop app to manage Agent Skills in one place and sync them to multiple AI coding tools’ global skills directories — “Install once, sync ever…

天枢 (Tianshu) 是一个基于harness工程的终端编程智能体运行时（Tui X Gui），针对DeepSeek V4 做了前缀缓存工程优化（长会话实测稳态命中率 97–99%）和深度适配。它跳出了传统 AI 编程助手把大模型仅当成“工具”的局限，基于认知虚拟机 (CVM)、自感知层和信息素（Stigmergy…