Field guide / sub-domain
29
sources in this field
Updated September 20, 2026
Current thesis
The shortest path to orientation.
This synthesis records claims and practices from the cited sources; reported outcomes and product capabilities have not been independently verified.
Skills distribution organizes markdown-based SKILL.md files (e.g., cold-email-copy) with self-improving loops: define criteria, run trials, score, rewrite, retest. Claude loads only relevant skills, keeping context small and preventing output drift. The AI PM OS centralizes 243 PM skills (v2.5, $499/year for up to 10 PMs), updated biweekly. Archify transforms codebases into typed JSON IR compiled into self-contained diagrams supporting five types with 8–12 core components; Architecture Delta mode compares snapshots for PR review. The 'show-me' skill instructs agents to pick the smallest visual format—pseudocode, call trees, component trees, file trees, Mermaid, or diffs—that clarifies key points. Diffs matched to topic shape typically outweigh full-block views unless content is mostly new. Matt Pocock notes this makes PR descriptions 'extremely easy to read.' Adoption mechanics now include: install the official Jev skill with npx skills add typesafe-ai/skills --skill typesafe-ai, then prompt your coding agent to analyze the project and identify where/how to apply it, offloading discovery work. Broader infrastructure remains constrained by model-guidance changes and lacks version control, ownership, searchability, and rigorous qualitative testing.
Evidence board
Claims worth carrying forward01Install the official Jev skill for existing projects with: `npx skills add typesafe-ai/skills --skill typesafe-ai`
02Workflow for adopting a new tool/skill in an existing codebase: install the skill, then prompt your coding agent to analyze the project and identify where/how to apply it—offloading discovery work to the agent rather than manual audit.
03Restructuring GTM around Claude Code as orchestrator (with tools like ColdIQ MCP as endpoints) cut a campaign build from two days of manual CSV-shuffling across four SaaS tools to one prompt and ~20 minutes of agent work.
04Project structure convention: CLAUDE.md (auto-loaded, under 200 lines, standing rules) is kept separate from brain.md (ICP/personas/scoring, referenced explicitly via @brain.md) so GTM substance doesn't bloat context on unrelated runs.
05brain.md should end with an explicit source-ranking block (e.g., closed-won CRM data > brain.md > last quarter's notes) so the model doesn't silently pick a winner when the ICP doc, intelligence store, and old notes contradict each other.
06MCP setup gotchas: unquoted env vars in `claude mcp add` cause shell expansion that leaks the live API key into .mcp.json; always write explicit "type": "stdio" or a url-only remote entry silently fails; .mcp.json is read only at session start.
07Never let Claude @-reference .env files—contents flow into the prompt, session transcript, and downstream outputs with no default block. Use a settings.json deny rule (Read(./.env) etc.) and hand the model a stripped .env.example instead.
08Seven-step outbound loop: detect signal → score/tier (1-100 via scoring.md) → find contacts → enrich/validate (email then phone, phone ~10x cost of email) → generate tiered copy → route to sequencer → analyze results back into the intelligence store.
Adjacent fields
Latest evidence
Bowe Frankema
Bowe Frankema shares a quick workflow for adopting 'Jev' (a typesafe-ai tool) in existing projects: install the official skill via CLI, then have an AI agent analyze the codebase for integration points.
Michel Lieben
A GTM agency's detailed playbook for running outbound campaigns through Claude Code as the orchestrator, with Apollo/Clay/Instantly-equivalent tools (via MCP) as endpoints. Covers folder structure, MCP connection setup, a seven-step signal-to-meeting workflow, skills files, reusable prompts, and eight recurring failure modes.
Matt Pocock
Humanlayer's 'show-me' skill (by @dexhorthy) is a markdown-defined toolbox of visual representations—pseudocode, call trees, component trees, file trees, Mermaid diagrams, diffs, and standalone HTML artifacts—used to make code and PR changes easier to understand at a glance.
Shann³
A promotional post introducing GPT-6 Astra with a benchmark claim and three concrete workflows for using it: auditing agent session history for automatable patterns, auditing a personal knowledge base/second brain, and reviewing stalled 'evergreen' projects with a plan-then-delegate execution model.
Vox
Archify is an open-source agent skill (29K+ GitHub stars) that lets Codex, Claude Code, Cursor, and OpenCode generate self-contained, verifiable HTML system diagrams (architecture, workflow, sequence, data-flow, lifecycle) from a codebase or description, with search, upstream/downstream tracing, and before/after comparison.
Charlie Hills
Anthropic's Claude Code team formally defines agentic 'loops' as cycles of work repeated until a stop condition is met, categorizing them into four types (turn-based, goal-based, time-based, proactive) by what gets handed off. Charlie Hills synthesizes the framework with practical commands (/goal, /loop, /schedule, skills) and a warning that every loop still needs an explicit stop condition or it burns tokens unattended.
Khe Hy
Khe Hy argues that individually-created AI skills (5-20 per person on buyside teams) work well but lack infrastructure for sharing, versioning, ownership, and testing—creating organizational duplication and drift. A quoted proposal suggests Notion could become a collaborative skills library to solve this.
George from 🕹prodmgmt.world
A PM skill (mckinsey-issue-tree, part of AI PM OS) applies Minto's Pyramid Principle and consulting-style problem trees (Why/What/How) so teams agree on what problem they're solving before jumping to causes, plans, or solutions, using AI to draft branches and run MECE checks repeatably.
Ahmad
A brief tip suggesting setting up an autoresearcher process that reviews past agent session traces to detect repeated patterns and behaviors, converting them into reusable skills.
Ben Holmes
Ben Holmes shares five custom agent skills/commands he uses to improve output quality from Claude Opus and GPT-5.6: /grill-me for research, /taste-review for design, /vercel-react-best-practices for React code quality, /simplify to cut fluff, and /test-app for end-to-end verification.