Sam's Skills
Sam's Skills contains reusable workflows for Codex, Claude Code, and other agents that support the Agent Skills format. Install a single skill or the collection bundle. Each skill keeps its instructions in SKILL.md and loads supporting references when needed.
Skills
| Skill | Use it for | Guide |
|---|---|---|
| Phone Size | Condense tasks, decisions, research, and updates into a phone-screen brief with supporting detail. | phone-size |
| Sam's Tree | Organize notes and other material into familiar, scannable trees. | sams-tree |
| Sam's Status | Show project work as a status tree with sources, progress, and missing information. | sams-status |
| Sam's Env | Draft environment examples with setup links, value guidance, annotation checks, and consistent spacing. | sams-env |
| Andrej Explain | Explain complex material and agent work through useful text, diagrams, HTML, or video. | andrej-explain |
| Migrate Custom GPTs | Back up GPT workflows privately and convert them to portable skills while tracking migration gaps. | migrate-cgpt |
| Sales Research | Research sales decisions and draft buyer communication using relevant practitioner sources. | sales-research |
| Sam's Gen Writing | Draft, review, or plan AI-assisted pages using Google's published Search guidance. | sams-gen-writing |
| Sam's Skill Creator | Create and improve agent skills with locked authoring sources and behavioral validation. | sams-skill-creator |
| Sam UX Audit | Audit websites and repository UX with locked review dependencies and evidence-backed findings. | sam-ux-audit |
| Zach Prompting | Improve prompts and agent instructions while preserving their intent and safeguards. | zach-prompting |
| HTMA Measure | Estimate uncertain quantities with ranges, sources, and decision recommendations. | htma-measure |
| Agent Verify | Verify new work and compare before/after improvements with fresh subagents and independent QA. | agent-verify |
Install a listed skill with npx skills add ThatGuySam/skills --skill <id>. Each linked guide includes usage examples and verification limits. Sam's Skill Creator and Sam UX Audit restore their locked dependencies on first use.
Try a skill
These examples show a few ways to start after installation:
Use $sams-skill-creator to turn this workflow into a focused, tested skill.
Use $sams-env to draft this project's .env.example.
Use $htma-measure to create a planning allowance for this project before quotes arrive.
For skill authoring, review the validation record. First use requires Node.js, npm, and Git to restore the private dependency set. Ordinary runs keep that set locked. Request a refresh explicitly.
For measurement, the HTMA Measure README explains the workflow, and its evaluation suite records verification. For environment templates, the synthetic test contract documents the checker limits.
Documentation
- Human guide: skills.samcarlton.com
- Complete agent corpus: skills.samcarlton.com/llms-full.txt
- Site source:
apps/docs
Other installation paths
Clone the repository and copy the required directory from skills/ into your agent's skills directory:
git clone https://github.com/thatguysam/skills.git
The repository also ships a marketplace bundle for Codex and Claude Code. The bundle loads every published skill; use the Skills CLI when you want only one. Its stable install ID remains htma-measure, while Claude Code uses the collection-level sam command namespace.
Codex
codex plugin marketplace add thatguysam/skills
codex plugin add htma-measure@thatguysam-skills
Claude Code
/plugin marketplace add thatguysam/skills
/plugin install htma-measure@thatguysam-skills
Claude Code namespaces bundled skills with sam, for example /sam:zach-prompting and /sam:htma-measure. Standalone installations keep their unbundled names. Codex skill selectors remain $zach-prompting and $htma-measure.
Repository structure
skills/
├── phone-size/
├── sams-tree/
├── sams-status/
├── sams-env/
├── andrej-explain/
├── migrate-cgpt/
├── sales-research/
├── sams-gen-writing/
├── sams-skill-creator/
├── sam-ux-audit/
├── zach-prompting/
├── htma-measure/
└── agent-verify/
Each skill keeps discovery metadata and the core workflow in SKILL.md, with conditional detail in supporting resources.
Acknowledgments
-
Andrej Explain is inspired by Andrej Karpathy’s writing about understanding model outputs and the sources in its research memo. It is not affiliated with or endorsed by him.
-
Sales Research uses original workflow instructions informed by the practitioners linked in its source guide. Their work remains with its publishers; inclusion does not imply endorsement.
-
Sam UX Audit composes the referenced skills from Addy Osmani, Emil Kowalski, and the community HIG reviewer by dickwu using Luisa Lima’s skills-lock. The HIG reviewer is not an Apple-authored package; each dependency retains its own provenance and terms.
-
Zach Prompting's name and original spark were inspired by Zach Miles. Its concrete prompting methods synthesize the cited OpenAI and Anthropic guidance; it is not affiliated with or endorsed by them.
-
HTMA Measure is inspired by measurement and uncertainty-management ideas popularized by Douglas W. Hubbard's How to Measure Anything. It is not affiliated with or endorsed by the author or publisher.
Contributing
Use trunk-based development for ordinary authorized changes: validate and publish directly to main, with branches or pull requests for explicit requests or required repository rules. Changes should keep each skill specific, portable, verifiable, and minimal. See CONTRIBUTING.md.
License
MIT. See LICENSE.
Sam's Skill Creator mixes methods from Anthropic, Matt Pocock, Poteto, and Zach Prompting using the locked upstream sources. Its bundled Codex creator snapshot retains its Apache-2.0 license; downloaded ingredients retain their respective licenses.