LLM Mart Basic

The greenfield decomposition interview. Routed to at startup when .codearbiter/CONTEXT.md lacks the <!--INITIALIZED--> body marker and no source code exists, or when the user invokes /decompose. A senior-architect persona drives a six-layer interview, persists every layer to disk

The parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units,

The checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint

The terminal step of /feature and /sprint. Routed to once commit-gate has cleared, to decide the branch's fate — merge via PR, open a PR, or discard. Direct merge to the default branch is forbidden; every change lands through a PR. Under /sprint the skill auto-selects "open PR" a

Finish an already-merged branch. Proves the branch is contained in the fetched default, classifies leftover artifacts as unique / redundant / superseded, returns to a clean --ff-only default checkout, and deletes the merged local branch — every discard confirmed per item. Routed

The behavior-preserving restructure gate. Routed to by /refactor for a rename, extract, inline, move, dedup, or internal-implementation swap. Six gated phases prove behavioral parity through unmodified pre-existing tests; any diff that classifies as `feat` is not a refactor and i

The single permitted path to a version tag, for any target declared in the project's .codearbiter/release-targets.md. Routed to when the user invokes /release on a non-default branch with a green suite. Takes the declared target as its one argument, derives the SemVer bump from C

The secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log

Optional, opt-in STRIDE threat pass for a sensitive feature — invoked deliberately via /threat-model, never forced on ordinary changes. Walks the change's attack surface and security boundaries (governed by {{PROJECT_DIR}}/.codearbiter/security-controls.md), surfaces threats and

The test-first gate. Routed to by /feature (after the spec is approved), /fix, and /refactor before any implementation code is written. Six gated phases — obligation scan, red, green, obligation verify, coverage, lint. No feature code exists before Phase 1 clears; nothing reaches

The deep, rarely-convened whole-codebase audit lane. Routed to when the user invokes {{CMD:tribunal}}. Seven gated phases — cost/model, map, roster dispatch, triage, report, approval+filing, telemetry. Costs on the order of millions of tokens; proceeds only after the user acknowl

OPTIONAL per-task isolation for autonomous parallel work. Routed to only on explicit opt-in by subagent-driven-development or dispatching-parallel-agents, so parallel units mutate files without colliding. Stands up one worktree per unit, works it in isolation, then integrates eac

The spec-to-plan bridge. Routed to by /feature once the brainstormed spec is approved, and by /sprint before execution. Decomposes the spec into 2–5 minute tasks, each carrying its exact file path(s) and a concrete verification step that maps to a tdd obligation. Writes the plan

The Socratic spec-refinement front of /feature, and the planning front of /sprint. Routed to BEFORE any code — it takes a one-line idea and drives it to an approved, concrete spec with testable acceptance criteria. Five gated phases — frame, shape, refine, write, review-and-appro

The only path to a commit. Routed to when the user invokes /commit or otherwise instructs codeArbiter to persist staged changes. Nine gated phases — permission, branch, classification, verification (test/lint/secrets), behavioral proof, diff review, selective stage, message, comm

Optional manual drift audit — report stale provenance-tracked docs (via _provenancelib drift detection across .codearbiter/.provenance/), then per stale doc offer re-scout / re-baseline / defer. Not the daily loop; commit-gate auto-heal owns routine maintenance.

The brownfield back-fill. Routed to by /create-context, and by startup when .codearbiter/CONTEXT.md lacks the <!--INITIALIZED--> body marker but source code exists. Six gated phases — pre-flight, scout dispatch, synthesis, gap interview, write, lock. Reads the existing codebase t

The banned-primitive gate. Routed to when changed code hashes, signs, encrypts, derives keys, generates security-relevant randomness, configures TLS, or imports a crypto library. Rejects broken primitives, disabled TLS verification, and home-rolled crypto; the approved-primitive

Investigate-then-decide root-cause analysis for a defect whose cause is unknown (distinct from /fix, which assumes a known bug). Five gated phases: capture, hypothesize, gather, decide, hand off. Investigation only, no code changes; exits to /fix, /adr, or a no-action close.

Author and track Architecture Decision Records. Routed to when the user invokes /adr to record a new decision or /adr-status to list ADR health. Authors numbered, dated, user-attributed ADRs under .codearbiter/decisions/, maintains supersede chains, and reports status read-only.

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.

Run autonomous hunt loop on a target — scope check → recon → rank surface → hunt → validate → report with configurable checkpoints. Usage: /autopilot target.com [--paranoid|--normal|--yolo]

Build an exploit chain — given bug A, finds B and C to combine for higher severity and payout. Knows common chain patterns: IDOR→ATO, SSRF→cloud metadata, XSS→ATO, open redirect→OAuth theft, S3→bundle→secret→OAuth. Usage: /chain

Active vulnerability hunting. Two-track dispatcher — asks Red Team vs WAPT, hands off to hunt-dispatch skill and sibling commands. Usage: /hunt target.com | /hunt *.target.com | /hunt targets.txt [--vuln-class X] [--source-code P] [--chrome]

On-demand intelligence fetch for a target — CVEs, disclosed reports, new features. Pulls NVD/GitHub-Advisory CVEs + bundled disclosed reports + hunt memory context. Usage: /intel target.com

Inspect or rotate the autopilot ledger JSONL files (findings.jsonl, negatives.jsonl). Caps file size and keeps N rotated backups so memory does not grow unbounded.

Pick up a previous hunt on a target — shows hunt history and untested surface from the autopilot ledger. Usage: /pickup target.com

Run full recon pipeline on a target — subdomain enum (Chaos API + subfinder), live host discovery (dnsx + httpx), URL crawl (katana + waybackurls + gau), gf pattern classification, nuclei scan. Outputs to recon/<target>/ directory. Usage: /recon target.com

Optional manual note on a target or the last confirmed finding. Capture is automatic during autopilot; this is for extra context. Usage: /remember

Write a submission-ready bug bounty report. Generates H1/Bugcrowd/Intigriti/Immunefi format with CVSS 3.1 score, proof of concept, impact statement, and remediation. Run /validate first. Usage: /report

Mandatory pre-flight scope check — verify an asset is in scope BEFORE any HTTP touch. Deterministic (deny-wins, default-deny) via engine/scope.py against the engagement's scope.md. Blocks out-of-scope testing. Usage: /scope <asset> [<asset> ...]

Show ranked attack surface for a target from its recon manifest + hunt memory. Deterministic backing is `cbh surface <target>` (reads recon/<target>/manifest.json); LLM layer adds ledger signal. Usage: /surface target.com

Meme coin and token security scan — checks for rug pull vectors (hidden mint, honeypot, fee manipulation, LP lock bypass, authority retention, bonding curve exploits, fake renounce, sandwich amplification). Manual 8-class grep audit (with an optional automated scanner if present). Usage: /token-scan <contract_path_or_dir> [--chain solana]

Quick 7-Question Gate triage on a finding before writing a report. Kills N/A submissions before they happen. Faster than /validate — for quick go/no-go decisions. Usage: /triage

Validate a finding — runs 7-Question Gate + 4-gate checklist. Kills weak findings before report writing. Prevents N/A submissions that hurt validity ratio. Usage: /validate

Smart contract security audit — runs through 10 bug class checklist (accounting desync, access control, incomplete path, off-by-one, oracle errors, ERC4626, reentrancy, flash loan, signature replay, proxy/upgrade). Applies pre-dive kill signals first. Generates Foundry PoC template for confirmed findings. Usage: /web3-audit <contract.sol>

Crabbox is a single CLI (`crabbox`). Commands are top-level, not nested under a

`crabbox actions` prepares a leased box from your repository's own GitHub

See [Runtime adapter stack](../features/runtime-adapter-stack.md) for the

`crabbox admin` groups trusted operator controls for coordinator-backed leases and the cloud resources behind them. Use it to inspect every lease the broker tracks, reconcile expired leases against live cloud state, force-release or delete a backing server, print provider IAM pol

`crabbox artifacts` turns a desktop lease into durable QA evidence: it collects

Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.

Agent OS: the agent gets smarter on its own. We just hold the line: the grading command and expected result never make it into the success contract we hand it.…

Curated systems, benchmarks, and papers etc. on memory for LLMs/MLLMs --- long-term context, retrieval, and reasoning.

:memo: Vimlike Modal Text Editor in Rust

CI-native security testing for MCP servers. Attack simulation, schema drift detection, and health scoring before agents depend on them.

Hermes Agent memory plugin/provider for scope-aware recall, SQLite truth, LanceDB semantic search, and hybrid retrieval.

For You Agent——AI 时代的个人随身数字人格。把你的模型、AI 账号、技能、提示词和工作方式，带到每一个 AI 工具里。

A coding agent: give it a prompt and it reads, writes, runs commands, and searches code in a loop until the work is done, using native tool-calling across OpenA…

A secure persistent personal agent server in Rust. One binary, sandboxed execution, multi-provider LLMs, voice, memory, Telegram, WhatsApp, Discord, Teams, and…

Self-evolving agent: grows skill tree from 3.3K-line seed, achieving full system control with 6x less token consumption

DeepSeek Harness Desktop (dsh-desktop). EAC: Embracing All Creation (揽尽万象). Bundled Node.js runtime with full dsh-CLI kernel, one-click startup, 10 built-in UI…

See your agent think. Zero-config observability & governance for 26 AI agent runtimes: Claude Code, Cursor, OpenAI Codex, GitHub Copilot, Gemini CLI, Cline, Ope…

Save 94% on AI coding tokens. Index your codebase, agents search instead of reading files. Works with Claude Code, Codex, Copilot, Cursor, Gemini CLI. Local MCP…

Yet another coding agent harness, lightweight and written in go.

a coding Agent from pi. ∞ providers, sub-agents, hashline edits, and a permission gate

Omnigent is an open-source AI agent framework and meta-harness: orchestrate Claude Code, Codex, Cursor, Pi, and custom agents — swap harnesses without rewriting…

🧠 Leon is your open-source personal assistant.

The Station, an open-world multi-agent environment that models a miniature scientific ecosystem.

The Frontend Stack for Agents & Generative UI. React, Angular, Mobile, Slack, and more. Makers of the AG-UI Protocol

VelaTerm = iTerm2 + Codex, The Best Terminal for AI Coding