{"id":47807011,"url":"https://github.com/cablate/ai-toolkit","last_synced_at":"2026-04-03T17:46:46.579Z","repository":{"id":345967395,"uuid":"1184154005","full_name":"cablate/ai-toolkit","owner":"cablate","description":"Battle-tested Claude Code skills for relentless AI delivery","archived":false,"fork":false,"pushed_at":"2026-04-03T05:40:19.000Z","size":384,"stargazers_count":4,"open_issues_count":5,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-04-03T12:56:03.927Z","etag":null,"topics":["ai-agents","ai-toolkit","claude-code","claude-code-skills","developer-tools","prompt-engineering","subagents"],"latest_commit_sha":null,"homepage":null,"language":"PowerShell","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/cablate.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-17T09:59:39.000Z","updated_at":"2026-04-03T05:40:22.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/cablate/ai-toolkit","commit_stats":null,"previous_names":["cablate/cablate-skills"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/cablate/ai-toolkit","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cablate%2Fai-toolkit","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cablate%2Fai-toolkit/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cablate%2Fai-toolkit/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cablate%2Fai-toolkit/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/cablate","download_url":"https://codeload.github.com/cablate/ai-toolkit/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cablate%2Fai-toolkit/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31366558,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-03T17:13:05.644Z","status":"ssl_error","status_checked_at":"2026-04-03T17:13:04.413Z","response_time":107,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-agents","ai-toolkit","claude-code","claude-code-skills","developer-tools","prompt-engineering","subagents"],"created_at":"2026-04-03T17:46:45.801Z","updated_at":"2026-04-03T17:46:46.571Z","avatar_url":"https://github.com/cablate.png","language":"PowerShell","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n  \u003cimg src=\"assets/banner.svg\" alt=\"AI Toolkit\" width=\"100%\"\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"README.zh-TW.md\"\u003e繁體中文\u003c/a\u003e\n\u003c/p\u003e\n\nA batteries-included working environment for [Claude Code](https://docs.anthropic.com/en/docs/claude-code). Dispatch agents, skills, and statusline — all extracted from daily production use.\n\n## What's Inside\n\n```\nai-toolkit/\n├── agents/                  # SOP agents (symlink to ~/.claude/agents/)\n│   ├── analyst.md           # Architecture / Planning / Audit\n│   ├── investigator.md      # Search / Explore / Debug / External research\n│   ├── builder.md           # Code implementation / Testing\n│   ├── reviewer.md          # Code review / Dead code cleanup\n│   ├── doc-sync.md          # Doc init / Doc sync\n│   └── agent-factory.md     # Design and generate new agents\n├── skills/                  # Skills (symlink to ~/.claude/skills/)\n│   ├── handoff/             # Session handoff\n│   ├── thorough/            # Relentless delivery mode\n│   ├── vector-memory/       # Persistent vector memory usage guide\n│   ├── project-docs/        # Project documentation structure\n│   ├── agentskill-expertise/ # Skill design knowledge base\n│   ├── collaboration-style/ # AI-human collaboration framework\n│   └── self-growth/         # Continuous learning framework\n├── mcp.example.json         # MCP server config template\n└── statusline/              # Cost \u0026 context monitoring\n```\n\n## Agents\n\nSOP-style prompts for Claude Code's Agent tool. When `/thorough` dispatches parallel subagents, prompt quality determines output quality — these agents provide step-based workflows with hard thresholds, classification heuristics, and structured output formats.\n\n| Agent | Model | When to use |\n|-------|-------|-------------|\n| [`analyst`](agents/analyst.md) | opus | \"design this\", \"plan the implementation\", \"audit codebase health\" |\n| [`investigator`](agents/investigator.md) | sonnet | \"find all usages of X\", \"how does this work\", \"why does this fail\" |\n| [`builder`](agents/builder.md) | sonnet | \"implement this\", \"modify the handler\", \"write tests for X\" |\n| [`reviewer`](agents/reviewer.md) | sonnet | \"review this code\", \"find dead code\", \"clean up unused exports\" |\n| [`doc-sync`](agents/doc-sync.md) | sonnet | \"set up project docs\", \"sync docs after changes\" |\n| [`agent-factory`](agents/agent-factory.md) | opus | \"create a new agent\", \"improve this agent's prompt\" |\n\nEach agent auto-detects its mode from dispatch context. One agent, multiple workflows.\n\n### Design Principles\n\n1. **Zero concept explanation** — All operational instructions. Claude already knows what CQRS is.\n2. **Step-based SOP** — \"Do X, then Y, if Z threshold → action.\" Not \"You are an expert at...\"\n3. **Hard rules as threshold + trigger** — `\u003e50 lines → flag`, `\u003e4 nesting levels → flag`. Not \"keep functions small.\"\n4. **Classification heuristics** — `AUTO-FIX / ASK / CRITICAL` with concrete criteria. Not checklists.\n5. **Structured output** — Every agent ends with a report template. Consistent, parseable.\n\n## Skills\n\n| Skill | Description |\n|-------|-------------|\n| [`/handoff`](skills/handoff/SKILL.md) | Session handoff — compress context into a structured prompt for seamless continuation |\n| [`/thorough`](skills/thorough/SKILL.md) | Relentless delivery mode — exhaust all options, cost-aware model selection, verify before done |\n| [`/vector-memory`](skills/vector-memory/SKILL.md) | Persistent vector memory via LanceDB — store facts, decisions, lessons across sessions |\n| [`/project-docs`](skills/project-docs/SKILL.md) | Project documentation structure — standard `proj-[name]/` layout with ADRs, stories, and operations guides |\n| [`/agentskill-expertise`](skills/agentskill-expertise/SKILL.md) | Agent Skill design knowledge base — mechanisms, philosophy, patterns, pitfalls |\n| [`/collaboration-style`](skills/collaboration-style/skill.md) | AI-human collaboration norms — friction cases, coding style, behavioral guidelines |\n| [`/self-growth`](skills/self-growth/SKILL.md) | Continuous learning framework — learn from work, organize knowledge, build feedback loops |\n\n## MCP Servers\n\nExample configuration for the MCP servers used in this toolkit.\n\n\u003e [`mcp.example.json`](mcp.example.json) — copy to your project as `.mcp.json` and fill in your API keys.\n\n| Server | What it does |\n|--------|-------------|\n| [`@cablate/memory-lancedb-mcp`](https://www.npmjs.com/package/@cablate/memory-lancedb-mcp) | Persistent vector memory with hybrid search (semantic + keyword) |\n| [Serena](https://github.com/oraios/serena) | Semantic code intelligence — symbol search, references, refactoring |\n\n## Statusline\n\nCost and context monitoring for Claude Code. Two-line display with context alerts and plan usage tracking.\n\n```\n Normal (\u003c 60% context):\n┌──────────────────────────────────────────────────────────────────┐\n│ Claude Opus 4  | [=======--------------] 45.2K/200.0K 22.6%    │\n│ 5h: 12.3% (4h 22m) | 7d: 8.1% (6d 3h)                        │\n└──────────────────────────────────────────────────────────────────┘\n\n Warning (\u003e= 60% context):\n┌──────────────────────────────────────────────────────────────────┐\n│ Claude Sonnet 4 | concise | [============--------] 130.5K/200.0K 65.3%  /handoff soon │\n│ 5h: 45.0% (2h 10m) | 7d: 22.4% (5d 1h)                       │\n└──────────────────────────────────────────────────────────────────┘\n\n Critical (\u003e= 80% context):\n┌──────────────────────────────────────────────────────────────────┐\n│ Claude Opus 4  | [==================--] 310.0K/200.0K 95.0%  !! HANDOFF NOW !! │\n│ 5h: 78.2% (1h 05m) | 7d: 51.3% (3d 12h)                      │\n│ !! DO NOT close/resume -- use /handoff first, or waste 6%+ of 5h tokens !! │\n└──────────────────────────────────────────────────────────────────┘\n```\n\n**Line 1** — Model name, output style (if not default), context progress bar with K-precision token counts, usage %, and alerts at 150K/200K/300K thresholds.\n\n**Line 2** — 5-hour and 7-day plan usage rates with reset countdowns. Fetched from Claude API (cached 5min) or inline `rate_limits` (v2.1.80+).\n\n**Line 3** — Appears at 250K+ tokens. Hard warning against closing/resuming without handoff.\n\n\u003e [`statusline/statusline.ps1`](statusline/statusline.ps1)\n\n```jsonc\n// ~/.claude/settings.json\n{ \"status_line_command\": \"powershell -NoProfile -File C:/Users/YOU/.claude/statusline.ps1\" }\n```\n\n## License\n\n[MIT](LICENSE)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcablate%2Fai-toolkit","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcablate%2Fai-toolkit","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcablate%2Fai-toolkit/lists"}