Standalone, AI-installable capabilities for Claude Code and other AI agent systems. Each pack is a self-contained prompt — point your DA at a directory and say "install this," and it sets everything up.
Two install paths exist:
- Individual skills — Every v5.0.0 skill ships as a standalone capability. Point your DA at the skill directory in the v5.0.0 release and ask it to install just that one.
- Curated bundles — Pre-built combinations with their own
INSTALL.mdandVERIFY.mdwizards (in this directory).
"Install the Research skill from PAI/Releases/v5.0.0/.claude/skills/Research/"
or for a curated bundle:
"Install the ContextSearch pack from PAI/Packs/ContextSearch/"
Your DA reads the contents 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