go-beast

Conducts structured technical conversations — code walkthroughs, architectural debates, decision support, rubber-duck debugging, and Q&A — using mode-aware dialogue, one-question-per-turn discipline, explicit assumption labeling, and escalation signals that hand off to the right beast when a conversation has produced enough to warrant a formal artifact.

Conducts a multi-pass code review with structured scoring across seven dimensions (correctness, security, maintainability, testability, performance, design, observability), applies a calibrated 0–4 rubric per dimension, produces OIR-structured findings with BLOCKER/WARNING/SUGGESTION/NIT severity classification, and emits a SCORE_REPORT.md with a dimensional scorecard and a merge verdict.

Translates approved requirements into an architecture decision record (ADR), technology stack selection, Mermaid component diagram, and interface contracts.

Audits an existing system's architecture across five dimensions — structure, observability, reliability, scalability, and security posture — and produces a prioritized HTML report with findings, capability gaps, and concrete improvement proposals. Uses repomix to pack the codebase, reads ADRs and CONTEXT.md when present, and opens the report in the browser.

Designs entity-relationship models, defines schemas with conventions, writes safe migrations with rollback, plans indexing strategy, and reviews queries for N+1 and sequential scans.

Translates interface contracts and functional requirements into BDD scenario files (Given/When/Then), an acceptance test skeleton, and a SPEC.md that go-wolf and go-lynx must satisfy before implementation begins.

Builds frontend UIs with correct component architecture, state management, API integration, accessibility, and responsive design — wired to a real backend, not mocked.

Designs and implements REST or GraphQL APIs, business logic layers, authentication, authorization, middleware, and server-side validation following a strict layered architecture.

Initializes go-beast explicitly for a new agent session or environment: classifies the target harness, chooses between planning-only/manual/installer-backed setup, separates canonical core steps from optional harness wiring, and validates that skills, instructions, and optional integrations are actually ready.

Profiles a running system to find real performance bottlenecks, establishes baseline measurements, applies targeted optimizations (query tuning, cache strategy, bundle reduction, async patterns), and validates improvement with before/after benchmarks. Produces PERF.md with evidence-backed findings and applied changes.

Runs a security review covering OWASP Top 10, authentication hardening, secrets management, dependency auditing, HTTP headers, and threat modeling for a software project.

Creates a working, runnable project skeleton — monorepo or multi-repo structure, dependency install, linter, formatter, Git hooks, env files, and dev server validation.

Designs and implements Workflow scripts (.js files that use agent(), pipeline(), parallel(), phase(), log(), and schema) for multi-agent orchestration. Covers: decomposing a task into phases, choosing pipeline vs parallel vs loop patterns, defining JSON schemas for structured agent output, writing the meta block, and registering the workflow in the pack.

Implements observability for a running system — structured logging with correlation IDs, metrics collection (Prometheus/OpenTelemetry), distributed tracing, health endpoints, alerting rules, and runbook-linked dashboards. Produces OBSERVABILITY.md documenting signal inventory, retention policy, and alert thresholds.

Designs the test pyramid, writes unit, integration, and end-to-end tests, establishes CI gates, and sets coverage policy for a software project.

Audits an existing go-* SKILL.md, identifies the specific weakness to fix (missing step, vague rule, incomplete output section, checklist gap, or eval-driven finding), applies the minimal edit, bumps the version, and verifies internal consistency. Never rewrites a working skill from scratch — every change targets exactly what is broken.

Conducts structured discovery interviews, produces a versioned REQUIREMENTS.md, identifies unknowns and risks, and generates a go-beast handoff plan for a software project.

Creates, audits, edits, and synchronizes AI agent context files — CLAUDE.md, AGENTS.md, GEMINI.md, COPILOT.md, CONTEXT.md, and agent memory files. Rewrites affected sections cleanly to improve agent behavior without causing regressions. Agent-agnostic: works with any context file format; Claude Code memory paths are one supported target among others.

Explores the solution space for approved requirements — generates 3–5 distinct approaches, evaluates each against the project's constraints, selects one with explicit rationale, and produces APPROACH.md as a decision record. Prevents go-fox from committing to an architecture before alternatives have been considered.

Plans, creates, validates, and cleans up isolated git worktrees for parallel development, risky refactors, or review staging; records provenance and safe cleanup rules.

Scans and reads all documentation in a project — README, CHANGELOG, CONTRIBUTING, CLAUDE.md, AGENTS.md, GEMINI.md, /docs, /wiki, and key config files — then produces a compact briefing that stays in session context for downstream tasks and other skills.

Audits and writes technical documentation for a software project — README, API reference, architecture docs, ADRs, runbooks, and changelog — ensuring every document is accurate, complete, and runnable.

Designs CI/CD pipelines, environment promotion strategy, infrastructure-as-code, secrets management in CI, and release automation for a software project.

Designs, writes, and validates new skills for the go-* family — from gap analysis to SKILL.md authoring, naming, description quality, workflow structure, and integration into the pack's handoff chain.

Designs, writes, tests, and registers lifecycle hooks for hook-capable coding agents — currently Claude Code and Codex. Produces hook scripts, wires them into the agent's hook configuration, and verifies execution.

Reviews a diff, branch, task, or named artifact against requirements, behavioral risk, security, and regression potential; produces severity-ranked findings, open questions, and a merge recommendation.

Designs and maintains Obsidian vaults — folder structure, naming conventions, wikilink strategy, plugin configuration (Dataview, Templater, Tasks, Kanban, etc.), note templates, and MOC (Map of Content) architecture. Produces a vault specification document, template files, and plugin config recommendations ready to apply.

Audits an existing lifecycle hook script for Claude Code or Codex, records the current contract (hook name, event, behaviour, exit codes), classifies the requested change (threshold / condition / behaviour / bug fix), applies the minimal edit as an explicit before/after diff, and re-validates with direct shell tests. Never rewrites a working hook from scratch — every change must be proven by running the hook with real JSON inputs.