4ge

Plugin 4ge from marketplace turd-box in turdpusher360/turd-box. Commands read from the repo's manifest — 23 days ago.

Skills (42)

• audit-bull

  Coverage-maximizing adversarial audit — volume over precision, every file, every pattern, every hunch

• audit-config

  Claude Code configuration audit — read-only health pass over agent frontmatter, hook wiring drift, slash-command registration, MCP server entries, memory setup. Audits what already exists; net-new builds belong to implement-feature, broken-hook repair to fix-hook.

• audit-implementation

  Implementation quality audit — codebase-wide static sweep for cyclomatic complexity, duplication, unused exports, stale dependencies, tech-debt inventory. Not a diff review (use review-code) and not branch cleanup (use repo-hygiene).

• audit-integration

  Integration audit — contract testing, service dependency mapping, failure mode analysis

• audit-panel

  Tactical multi-angle parallel review of a target (diff range, file, or topic). Dispatches 2-4 reviewers with different perspectives (judgment, nitpick, DFE adversarial) and converges into a single verdict with agreement matrix. Use when: the user says 'review this', 'audit panel', 'get a second opin

• audit-security

  Comprehensive whole-repo security audit — OWASP posture, supply-chain integrity, Docker CIS benchmark, MCP attack surface, credential exposure. For a per-PR security pass use review-security.

• cc-update

  Update cc, cc upgrade, new version, version audit, claude code update, CC upgraded, harness update. Use when Claude Code has been updated to a new version and the codebase needs to be audited for impact.

• debug-investigate

  Systematic bug investigation — error diagnosis, root cause analysis, fix verification

• design-suite

  Contextual design assistance — auto-detects design mode and assembles appropriate toolkit

• dfe-review

  Heavyweight 6-pass adversarial review — fans out to 5 domain-minion subagents + 1 synthesis pass for deep/risky changes. For a quick single-pass review use review-adversarial.

• explain-codebase

  Codebase explanation and onboarding — architecture walkthroughs, how-does-X-work answers

• fix-aisle

  AISLE security posture fixes — prompt guard, scanner-cache readiness, shelved gate docs, and threat-management workflows

• fix-cloudflare

  Cloudflare Workers fixes — Hono routes, Durable Objects, WebSocket hub, webhooks, wrangler deploys

• fix-commander

  Electron tray app fixes — React 19 panels, Zustand stores, IPC bridge, Vite 8 build

• fix-d365

  Dynamics 365 Web API fixes — OAuth2, OData queries, entity schemas, webhook HMAC, D365 mock testing

• fix-hook

  Hook system bug fixing and development — hook-utils, settings.json wiring, stdin protocol, exit codes, capability patterns

• fix-hud

  HUD engine rendering fixes — zones, expressions, themes, badges, canvas, boot screen, statusline

• fix-kernel

  Agentic OS kernel fixes — boot sequence, capability registry, process registry, scheduler, services

• fix-plugin

  4ge plugin framework fixes — config system, dialect detection, trust scoring, telemetry, session archaeology

• fix-teams

  Microsoft Teams integration fixes — Bot Framework, Adaptive Card design + snapshot testing, Teams manifest

• fix-wizard

  Debug the wizard/outhouse engine source when /outhouse or /wizard itself misbehaves — stage-transition defects, threshold math, score-merge bugs, scan-mode detection. To run the pipeline use wizard-engine.

• forge

  Multi-teammate orchestrator — 7 phases: scope, brainstorm, spec, plan, execute, integrate, ship

• implement-feature

  Build new cross-cutting functionality spanning several surfaces at once — author a hook plus its settings.json entry, add a slash command, write supporting rules together. Greenfield multi-domain work; single-domain bug fixes route to the matching fix-* skill.

• ops-infra

  Infrastructure operations — health checks, deploys, log analysis, MCP diagnostics, Docker, monitoring

• plan-architecture
• prime
• reconcile
• repo-hygiene
• repo-onboard
• research-background
• research-multi
• research-single
• respawn
• review-adversarial
• review-code
• review-project
• review-security
• run-test
• session-audit
• session-tasking
• wizard-engine
• write-test

Slash commands (39)

• /4ge:4ge

  4ge ecosystem — forge orchestration + OS operations

• /4ge:aisle

  [AISLE] AI Security Learning Environment -- security posture, scanning, and threat management

• /4ge:audit

  Run code quality audit — 70 checks across 10 domains

• /4ge:autoresearch

  Autoresearch measurement plane — read-only status and scans of the repo's measures. Use: /autoresearch status, /autoresearch scan, /autoresearch <domain>

• /4ge:blueprint

  Bootstrap or update a Claude Code environment

• /4ge:constraint

  Log a constraint/dead-end to the DCD enrichment file. Feeds into /respawn context extraction.

• /4ge:debug

  Systematic debugging -- guided root-cause analysis with memory search, causal mapping, and fix verification

• /4ge:decide

  Log a decision to the DCD enrichment file. Feeds into /respawn context extraction.

• /4ge:design

  Contextual design assistant — auto-detects design mode (Visual/API/Data/System) and assembles the right toolkit

• /4ge:dfe

  AI code review — 6-pass DFE analysis (5 domain minions + 1 adversarial pass). Use /dfe [file|all] to catch hallucinated APIs, logic bugs, security holes, and runtime failures.

• /4ge:evolve

  Analyze usage telemetry and suggest config changes to evolve your .4ge/config.json

• /4ge:export

  Export session work as a business-ready deliverable (brief, deck, handoff)

• /4ge:fix

  Inline issue collector. Capture maintenance issues without interrupting workflow. Issues appear in the next /outhouse run.

• /4ge:forge

  Multi-teammate orchestrator. Use /forge <task>, /forge resume, or /forge park.

• /4ge:help

  4ge plugin glossary, command index, and usage guide. Use: /help for full index, /help <command> for command details, /help hooks, /help skills. Use when asking 'what commands are available', 'how do I use forge', 'what does this plugin do', or 'show me the 4ge commands'.

• /4ge:hitchhiker

  [REDIRECT] Use /recall instead. Knowledge search moved to the Guided Knowledge hub.

• /4ge:hud

  Toggle, configure, or route the OS HUD surfaces.

• /4ge:infra

  Docker container health monitoring. Use: /infra or /infra check to see container health, /infra heal <container> to restart a container

• /4ge:lint

  Show rule follow-through rates and suggest demotions for under-followed rules

• /4ge:lounge

  [HIDDEN] Lounge mode -- use --lounge flag on any command instead

• /4ge:maintain

  Redirects to /outhouse. Use /outhouse for repository maintenance.

• /4ge:onboard

  Repository onboarding scanner — maps the stack, flags missing CI/test/lint config, generates a setup checklist

• /4ge:outhouse

  Repository maintenance hub: health wizard, fix-collector, evolve, lint, blueprint, wizard suite. Use /outhouse for the menu or /outhouse --quick for a fast score.

• /4ge:pr

  Verify, commit, push, and open a PR

• /4ge:ps

  Agent process dashboard — running agents, zombie candidates, reaper log, and session uptime

• /4ge:recall

  Guided Knowledge hub: memory search, repo map, context budget, respawn, decisions. Use /recall for the menu or /recall <query> for dev-memory search.

• /4ge:recompile

  Recompile the task state from the user's recent messages: separate instruction / diagnosis / exploration / emphasis, return a compact task lock, and continue. Use when the conversation has drifted or after a heated correction.

• /4ge:reconcile

  Fold handoffs since the last reconcile into BACKLOG.md — re-surface gold mines, flag drift, archive stale _runs/, regenerate INDEX.md. Run every 5 sessions or on pivot/release.

• /4ge:releases

  Show recent release notes from shipped sessions

• /4ge:respawn

  Extract decision chain and prepare a fresh Claude instance with full context

• /4ge:secret

  Capture a secret to .env (or --file <path.env>) without crossing model context

• /4ge:ship

  Guided delivery hub: commit, push, PR, release notes, export, signoff. Use /ship for the menu or /ship <message> to commit+push directly.

• /4ge:signoff

  Review and enrich the session cartridge before ending. Writes a model-curated summary for the next session.

• /4ge:studio

  Activate Studio Mode — full HUD engine + reactive hooks + badge tracking

• /4ge:substrate

  Substrate composition engine — render text and OS state through Unicode combining marks, Math Alphanumerics, and block elements. Modes: os, max, palimpsest, alphabets, enclosed, help.

• /4ge:superdupersecret

  Alias of /secret - capture a secret to .env (or --file <path.env>) without crossing model context

• /4ge:tour

  5-step guided walkthrough of the highest-value /4ge commands. Use this when a user is new to 4ge, asks 'where do I start', or wants to see what the plugin can do. Trigger phrases: 'tour', 'getting started', 'show me how to use 4ge', 'what should I try first'.

• /4ge:usage

  Local usage/cost analytics from Claude Code transcripts — 5h billing blocks, burn rate, daily/monthly/session rollups, pre-dispatch gate

• /4ge:wizard

  Wizard suite launcher. Run /wizard list to see available modes, /wizard <mode> to launch a specific wizard.

Subagents (17)

• DFE

  Heavyweight 6-pass adversarial review SYNTHESIS agent (Dumb-Fuck Expert). Carries NO Agent tool and cannot fan out — the /4ge:dfe command fans the 5 domain minions out from the lead, then dispatches DFE to verify, deduplicate, and rule. Dispatched bare it runs all six passes in one context and loses

• dfe-artifacts

  DFE minion — Pass 6 (ARTIFACTS). Dead exports, orphaned variables, copy-paste drift, excessive complexity, TODO/FIXME without tickets.

• dfe-existence

  DFE minion — Pass 1 (EXISTENCE). Verifies imports resolve, packages exist on npm, APIs are not deprecated or hallucinated.

• dfe-logic

  DFE minion — Pass 3 (LOGIC). Race conditions, off-by-one, inverted booleans, tautological tests, async errors, falsy-value bugs.

• dfe-runtime

  DFE minion — Pass 4 (RUNTIME) + Pass 5 (TRUST). Environment mismatches, missing await, global state, boundary validation, type assertion hiding.

• dfe-security

  DFE minion — Pass 2 (SECURITY) + Pass 7 (PROVENANCE). Injection vectors, taint analysis, OWASP Top 10, secrets in source, prompt injection artifacts.

• forge-brainstorm

  Brainstorming phase subagent for forge sessions. Runs Phase 2 (brainstorm) and Phase 3 (spec writing) with the brainstorming skill injected. Reduces lead context burden by keeping skill content in subagent window.

• forge-planner

  Plan writing subagent for forge sessions. Runs Phase 4 (plan writing) with the writing-plans skill injected. Reads approved spec, produces DAG-structured execution plan.

• forge-shipper

  Ship and handoff subagent for forge sessions. Runs Phase 7 (verification, branch completion, triple-write handoff) with verification and finishing skills injected.

• general-purpose

  General-purpose agent — inherits parent model. Approved for A/B testing and ad-hoc tasks.

• master-auditor-46

  Legacy master-auditor compatibility variant; inherits runtime model unless explicitly overridden

• master-auditor

  Audit team lead - coordinates domain auditors, cross-correlates findings, produces MASTER-VERDICT with 100+ item checklist

• opus-audit

  Write-to-disk audit agent. Produces _runs/*.md reports per audit skill.

• opus-planner

  Architectural planner. Phased breakdowns, STOP gates, judgment calls.

• opus-review

  Read-only judgment agent. Code review, audit analysis, architectural critique.

• sonnet-execute

  General-purpose execution runtime for skill-dispatched work. All domain fix-* and ops-infra skills route here.

• sonnet-research

  Web research and documentation agent. Writes reports to _runs/.

Hooks (25)

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/hooks/sync-to-data.cjs"

• SessionStart

  node "${CLAUDE_PLUGIN_DATA}/hooks/os-boot.cjs"

• SessionStart

  node "${CLAUDE_PLUGIN_DATA}/hooks/file-integrity-guard.cjs" boot

• SessionStart

  node "${CLAUDE_PLUGIN_DATA}/hooks/whats-new.cjs"

• SessionStart

  node "${CLAUDE_PLUGIN_DATA}/hooks/telemetry-session.cjs"

• SessionStart

  node "${CLAUDE_PLUGIN_DATA}/hooks/license-refresh.cjs"

• Stop

  node "${CLAUDE_PLUGIN_DATA}/hooks/telemetry-session.cjs"

• Stop

  node "${CLAUDE_PLUGIN_DATA}/hooks/checkpoint-buddy.cjs"

• Stop

  node "${CLAUDE_PLUGIN_DATA}/hooks/hud-state-clear.cjs"

• PreToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/forge-prompt-lint.cjs"

• PreToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/sonnet-overflow-guard.cjs"

• PreToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/forge-scope-check.cjs"

• PreToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/weasley-conflict-check.cjs"

• PostToolUseFailure

  node "${CLAUDE_PLUGIN_DATA}/hooks/rubber-duck-debugger.cjs"

• PostToolUseFailure

  node "${CLAUDE_PLUGIN_DATA}/hooks/feedback-queue-capture.cjs"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_DATA}/hooks/secret-redact.cjs"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_DATA}/hooks/prompt-scaffolder.cjs"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_DATA}/hooks/hud-message-refresh.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/forge-heartbeat.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/os-accounting.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/weasley-heartbeat.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/file-integrity-guard.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/dfe-post-edit.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/file-integrity-guard.cjs"

• PostToolUse

  node "${CLAUDE_PLUGIN_DATA}/hooks/hud-reactive.cjs"

Similar plugins