ngxccc/agent-skills-kit
Your AI forgets. This remembers. Spec-driven coding harness for vibecoders, product owners, CEOs and real builders — self-improving context memory. Kills context rot, ships features, not spaghetti. Claude Code & Codex. Any stack. 30 seconds
ARCHETYPE
Fort Knox
Deny lists, pre-tool hooks, sandboxing. Nothing touches prod without a signature.
GUARDRAILS
4/5
Blocks destructive commands · Pre-tool screening hook · No YOLO mode · Sandbox or ask-first rules · details
Copy this rig
# review before running: this installs third-party code $ claude mcp add --transport http mdn https://mcp.mdn.mozilla.net/ $ claude mcp add chrome-devtools -- bunx -y chrome-devtools-mcp@latest $ claude mcp add playwright -- bunx -y @playwright/mcp@latest $ claude mcp add codebase-memory -- codebase-memory-mcp server $ npx degit ngxccc/agent-skills-kit/.claude ./rig-agent-skills-kit # inspect, then merge into .claude/
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add --transport http mdn https://mcp.mdn.mozilla.net/ $ claude mcp add chrome-devtools -- bunx -y chrome-devtools-mcp@latest $ claude mcp add playwright -- bunx -y @playwright/mcp@latest $ claude mcp add codebase-memory -- codebase-memory-mcp server
$ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-agent-browser .claude/skills/ag-agent-browser $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-agent-strategy-compare .claude/skills/ag-agent-strategy-compare $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-architect-verifier .claude/skills/ag-architect-verifier $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-audit-ag .claude/skills/ag-audit-ag $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-audit-context .claude/skills/ag-audit-context $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-audit-plans .claude/skills/ag-audit-plans $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-autoresearch .claude/skills/ag-autoresearch $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-brainstorming .claude/skills/ag-brainstorming $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-code-interrogation .claude/skills/ag-code-interrogation $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-context-discovery .claude/skills/ag-context-discovery $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-debug .claude/skills/ag-debug $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-docs-seeker .claude/skills/ag-docs-seeker $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-docs .claude/skills/ag-docs $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-feasibility-test .claude/skills/ag-feasibility-test $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-frontend-design .claude/skills/ag-frontend-design $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-generate-closeout .claude/skills/ag-generate-closeout $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-generate-context .claude/skills/ag-generate-context $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-generate-plan .claude/skills/ag-generate-plan $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-generate-spec .claude/skills/ag-generate-spec $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-git-flow .claude/skills/ag-git-flow $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-harness-sync .claude/skills/ag-harness-sync $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-intent-clarify .claude/skills/ag-intent-clarify $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-mcp-management .claude/skills/ag-mcp-management $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-plan-discovery .claude/skills/ag-plan-discovery $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-predict .claude/skills/ag-predict $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-preview .claude/skills/ag-preview $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-problem-solving .claude/skills/ag-problem-solving $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-red-team-plan .claude/skills/ag-red-team-plan $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-repomix .claude/skills/ag-repomix $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-reverse-code-review .claude/skills/ag-reverse-code-review $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-risk-evidence-pack .claude/skills/ag-risk-evidence-pack $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-scenario .claude/skills/ag-scenario $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-scout .claude/skills/ag-scout $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-second-brain .claude/skills/ag-second-brain $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-security .claude/skills/ag-security $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-sequential-thinking .claude/skills/ag-sequential-thinking $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-setup .claude/skills/ag-setup $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-socratic-mentor .claude/skills/ag-socratic-mentor $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-strict-config-derivation .claude/skills/ag-strict-config-derivation $ npx degit ngxccc/agent-skills-kit/.claude/skills/ag-team .claude/skills/ag-team
$ curl -fsSL --create-dirs -o .claude/agents/ag-agent-browser.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-agent-browser.md $ curl -fsSL --create-dirs -o .claude/agents/ag-code-reviewer.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/ag-code-simplifier.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-code-simplifier.md $ curl -fsSL --create-dirs -o .claude/agents/ag-debugger.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-debugger.md $ curl -fsSL --create-dirs -o .claude/agents/ag-execute-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-execute-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-fast-mode-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-fast-mode-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-git-manager.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-git-manager.md $ curl -fsSL --create-dirs -o .claude/agents/ag-innovate-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-innovate-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-plan-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-plan-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-quick-fix-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-quick-fix-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-research-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-research-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-security-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-security-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-spec-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-spec-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-tester.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-tester.md $ curl -fsSL --create-dirs -o .claude/agents/ag-ui-ux-designer.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-ui-ux-designer.md $ curl -fsSL --create-dirs -o .claude/agents/ag-update-process-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-update-process-agent.md $ curl -fsSL --create-dirs -o .claude/agents/ag-validate-agent.md https://raw.githubusercontent.com/ngxccc/agent-skills-kit/main/.claude/agents/ag-validate-agent.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Write",
"hooks": [
{
"type": "command",
"command": "node \"$CLAUDE_PROJECT_DIR\"/.claude/hooks/descriptive-name.cjs"
}
]
},
{
"matcher": "Bash|Glob|Grep|Read|Edit|Write",
"hooks": [
{
"type": "command",
"command": "node \"$CLAUDE_PROJECT_DIR\"/.claude/hooks/scout-block.cjs"
},
{
"type": "command",
"command": "node \"$CLAUDE_PROJECT_DIR\"/.claude/hooks/privacy-block.cjs"
}
]
}
]
}
} MCP servers (4)
| server | source | est. tokens |
|---|---|---|
| mcp.mdn.mozilla.net · "mdn" | remote · remote | 2.5k |
| Chrome DevTools MCP · "chrome-devtools" | npm | 9.0k |
| Playwright MCP · "playwright" | npm | 7.5k |
| codebase-memory-mcp · "codebase-memory" | binary | 2.5k |
Skills (46)
ag-agent-browserag-agent-strategy-compareag-architect-verifierag-audit-agag-audit-contextag-audit-plansag-autoresearchag-brainstormingag-code-interrogationag-context-discoveryag-debugag-docs-seekerag-docsag-feasibility-testag-frontend-designag-generate-closeoutag-generate-contextag-generate-planag-generate-specag-git-flowag-harness-syncag-intent-clarifyag-mcp-managementag-plan-discoveryag-predictag-previewag-problem-solvingag-red-team-planag-repomixag-reverse-code-reviewag-risk-evidence-packag-scenarioag-scoutag-second-brainag-securityag-sequential-thinkingag-setupag-socratic-mentorag-strict-config-derivationag-teamag-tech-graphag-test-coverage-planag-validate-findingsag-watzupag-web-testingag-xia
Subagents (17)
| ag-agent-browser model: google-antigravity/gemini-3.6-flash | Specialist AI browser automation agent. Executes long autonomous browser sessions, visual DOM snapshotting, Puppeteer/DevTools script automation, E2E testing, visual regression analysis, and cloud bro |
| ag-code-reviewer model: sonnet | Comprehensive code review with scout-based edge case detection. Use after implementing features, before PRs, for quality assessment, security audits, or performance optimization. |
| ag-code-simplifier model: sonnet | Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Focuses on recently modified code unless instructed otherwise. |
| ag-debugger model: sonnet | Investigate issues, analyze system behavior, diagnose performance problems, examine database structures, collect logs, run debugging tests, or optimize performance. |
| ag-execute-agent model: opus | EXECUTE MODE - Implementing EXACTLY what was planned. Full tool access. Can only be invoked after explicit user confirmation. Use after plan is approved. |
| ag-fast-mode-agent model: opus | FAST MODE - Execute compressed RIPER-5 workflow (RESEARCH + SPEC + INNOVATE + PLAN + VALIDATE) in one session, then pause for EXECUTE confirmation. Use when you want quick end-to-end solution. |
| ag-git-manager model: sonnet | Stage, commit, and push code changes with conventional commits. Use when user says "commit", "push", or finishes a feature/fix. |
| ag-innovate-agent model: opus | INNOVATE MODE - Brainstorming and option exploration. Use for discussing implementation approaches, architectural trade-offs, and technical designs. Never writes plan files or code. |
| ag-plan-agent model: opus | PLAN MODE - Creating exhaustive technical specifications and implementation plans. Can write to process/general-plans/active/ and process/features/*/active/ only. Use after approach is decided. |
| ag-quick-fix-agent model: opus | Quick-fix lane for small low-risk changes. Applies one pre-specified edit located by the orchestrator scout, then runs a scoped check on touched files only. No plan file, no validate-contract, no EVL, |
| ag-research-agent model: opus | RESEARCH MODE - Information gathering only. Use for understanding existing code, architecture, and context. Never suggests implementations or modifications. |
| security-agent model: google-antigravity/gemini-3.6-flash | Dedicated SAST & Security Auditor specializing in OWASP Top 10/ASVS, STRIDE Threat Modeling, Zero-Day logic flaw detection, and Auth/Cryptographic boundary verification. |
| ag-spec-agent model: sonnet | SPEC MODE - Product-discovery requirements doc for user review. Use after RESEARCH, before INNOVATE, to turn research findings plus user intent into a reviewable requirements artifact (user stories, a |
| ag-tester model: opus | Validate code quality through testing: unit/integration tests, test coverage analysis, error handling, performance requirements, and build verification. Call after implementing features or significant |
| ag-ui-ux-designer model: sonnet | UI/UX execution support including interface implementation, design-system polish, responsive layouts, micro-animations, accessibility review, or design documentation. |
| ag-update-process-agent model: opus | UPDATE PROCESS MODE - Documenting learnings, updating rules, archiving completed plans, and maintaining repo context. Use after feature implementation is complete. |
| ag-validate-agent model: opus | VALIDATE MODE - Two-layer feasibility and validation fan-out to verify implementation plans before execution. Writes ## Validate Contract into plan file. Use after PLAN mode. |
Hooks (10)
| event | matcher | runs |
|---|---|---|
| SessionStart | * | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/session-init.cjs |
| SubagentStart | * | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/subagent-init.cjs |
| PreToolUse | Write | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/descriptive-name.cjs |
| PreToolUse | Bash|Glob|Grep|Read|Edit|Write | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/scout-block.cjs |
| PreToolUse | Bash|Glob|Grep|Read|Edit|Write | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/privacy-block.cjs |
| PostToolUse | Edit|Write|MultiEdit | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/session-state.cjs |
| PostToolUse | Edit|Write|MultiEdit | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/post-edit-simplify-reminder.cjs |
| PostToolUse | Write | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/post-write-plan-check.mjs |
| PostToolUse | Bash | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/post-commit-lint.mjs |
| Stop | * | node "$CLAUDE_PROJECT_DIR"/.claude/hooks/stop-validator-sweep.cjs |
Similar rigs
firu-daniel/autonomous-sdlc-harness
An autonomous software-delivery harness: a task prompt becomes a planned, implemented, reviewed and pushed branch, driven by a fleet of single-purpose agents.
Pragmatist 16.5k tok ·
terror/dotfiles
My scripts and configurations
Pragmatist 25.0k tok ·
nextlevelbuilder/ui-ux-pro-max-skill
An AI skill that provides design intelligence for building professional UI/UX across multiple platforms.
Pragmatist 21.5k tok ·
tdimino/claude-code-minoan
A curated ~/.claude/ configuration for professional development workflows — 90+ skills, 46 hooks, and CLI tools
MCP Hoarder 55.3k tok ·