Standalone, AI-installable capabilities for Claude Code and other AI agent systems. Each pack is a self-contained directory — point your DA at it and say "install this," and it sets everything up.
Every v5.0.0 skill ships as its own pack below. Each pack contains README.md, INSTALL.md, VERIFY.md, and src/ (the actual skill source).
"Install the Research pack from PAI/Packs/Research/"
Your DA reads INSTALL.md and walks through a 5-phase wizard: system analysis, user questions, backup, installation, verification.
| Skill | What it does |
|---|---|
| Agents | Compose custom agents from base traits, voice, and specialization; manage functional teams |
| ApertureOscillation | 3-pass scope oscillation: narrow tactical, wide strategic, then synthesis to surface design tensions |
| Aphorisms | Curated aphorism collection with content matching, themed search, thinker research |
| Apify | Scrape Instagram, LinkedIn, TikTok, YouTube, Facebook, Google Maps, e-commerce via Apify actors |
| Art | Static visual content across 20+ formats via Flux, Nano Banana Pro, GPT-Image-1 |
| ArXiv | Search and retrieve arXiv academic papers with AlphaXiv-enriched AI summaries |
| AudioEditor | Whisper transcription → Claude classification → ffmpeg cuts for audio/video editing |
| BeCreative | Divergent ideation via Verbalized Sampling and extended thinking |
| BitterPillEngineering | Audit AI instruction sets for over-prompting; classify rules CUT/RESOLVE/MERGE/KEEP |
| BrightData | 4-tier progressive scraping with auto-escalation through WebFetch, curl, browser, MCP proxy |
| Browser | Headless browser automation via agent-browser Rust CLI for batch and parallel work |
| ContextSearch | 2-phase context search across session registry, work directories, and ISAs for cold-start recovery |
| Council | Multi-agent collaborative debate with round-by-round transcripts and intellectual friction |
| CreateCLI | Generate production TypeScript CLIs via 3-tier template system (manual, Commander, oclif) |
| CreateSkill | Complete PAI skill development lifecycle — scaffold, validate, canonicalize, evaluate |
| Daemon | Manage public daemon profile — living digital presence with deterministic security filtering |
| Delegation | Six parallelization patterns: built-in agents, worktrees, background tasks, custom agents |
| Evals | Agent evaluation framework with code-based, model-based, and human graders; pass@k scoring |
| ExtractWisdom | Content-adaptive wisdom extraction that detects domains and builds custom sections |
| Fabric | Execute any of 240+ Fabric prompt patterns natively across extraction, analysis, creation |
| FirstPrinciples | Physics-based reasoning that deconstructs problems to irreducible truths |
| Ideate | Evolutionary ideation engine — 9-phase loop for novel solution generation |
| Interceptor | Real Chrome browser automation with zero CDP fingerprint; mandatory for visual verification |
| Interview | Phased conversational interview across all PAI context files (TELOS first) |
| ISA | Owns the Ideal State Artifact primitive — articulate "done" for any kind of work |
| IterativeDepth | Multi-angle exploration through 2-8 sequential passes from different scientific lenses |
| Knowledge | Manage typed Knowledge Archive across People, Companies, Ideas, Research with 8 link types |
| Loop | Iterative improvement loop — refine a target across multiple Algorithm cycles |
| Migrate | Intake external content (Obsidian, Notion, Apple Notes, journals) and classify into PAI taxonomy |
| Optimize | Autonomous optimization loop — hill-climb any target with metrics or LLM-as-judge eval |
| PAIUpgrade | Generate prioritized PAI upgrade recommendations via 4 parallel research threads |
| PrivateInvestigator | Ethical people-finding and identity verification via 15 parallel research agents |
| Prompting | Meta-prompting standard library — generate, optimize, and compose prompts programmatically |
| RedTeam | Adversarial analysis via 32 parallel expert agents to stress-test ideas, strategies, plans |
| Remotion | Programmatic video with React via Remotion — compositions, sequences, motion graphics to MP4 |
| Research | Multi-agent research with 4 depth modes (quick, standard, extensive, deep investigation) |
| RootCauseAnalysis | Structured incident investigation: 5 Whys, Fishbone, Postmortem, Fault Tree, Kepner-Tregoe |
| Sales | Transform product docs into sales narratives with charcoal sketch art and talking points |
| Science | Scientific method as a universal problem-solving algorithm — goal, hypotheses, experiments |
| SystemsThinking | Structural analysis via Iceberg, Causal Loop, Archetypes, Leverage Points, Concept Map |
| Telos | Dual-context Life OS — read and update goals, beliefs, wisdom, missions, mental models |
| USMetrics | 68 US economic indicators from FRED, EIA, Treasury, BLS, Census APIs |
| Webdesign | Web/UI design via Anthropic's Claude Design (claude.ai/design) with frontend handoff |
| WorldThreatModel | Stress-test ideas, strategies, investments against 11 time horizons from 6 months to 50 years |
| WriteStory | Fiction across seven narrative layers — Storr, Pressfield, Forsyth |
Pre-built combinations with their own install wizards. Useful when you want a thematic grouping installed in one shot.
| Bundle | What's inside |
|---|---|
| ContextSearch | /context-search and /cs slash commands |
| Agents | Custom agent composition |
| ContentAnalysis | Wisdom extraction from videos, podcasts, articles, YouTube |
| Investigation | People search and identity verification |
| Media | Visual and video content — illustrations, diagrams, Remotion |
| Research | Multi-agent research with quick/standard/extensive/deep modes |
| Scraping | Bright Data + Apify scraping bundle |
| Telos | Life OS goals, beliefs, wisdom, dashboards |
| Thinking | First principles, council, red team, science, brainstorming |
| USMetrics | 68 US economic indicators |
| Utilities | Developer tools — CLI generation, skill scaffolding, Fabric, browser automation |
PackName/
├── README.md # What it does and why
├── INSTALL.md # Step-by-step wizard for AI-assisted installation
├── VERIFY.md # Post-install verification checklist
└── src/ # Source files to copy