Agentic Workspaces

Browse / General-purpose / my-ai-workflows

my-ai-workflows

PhantatP

Portable workflow configuration for Claude Code and Codex.

Open PhantatP/my-ai-workflows → Does not match the four rules.

Works with any agent

ships AGENTS.md, the cross-tool instruction file.

In its own words

README.mdPhantatP/my-ai-workflows

Portable workflow configuration for Claude Code and Codex. V2.2.1 adds deterministic BUILD enforcement: executable trigger scanning, fail-closed sensitive-command guarding, and a small human-readable operational log.

Contents

  • workflow/ — platform-neutral policy, executable trigger scanner, enforcement

adapters, minimal logging guidance, roles, and skill audit.

  • CLAUDE.md — thin Claude Code adapter for the shared policy.
  • AGENTS.md — thin Codex adapter for the shared policy.
  • skills/ — active V2.2 methodologies: BUILD routing, testing, and

independent falsification review.

  • archive/v1/ — remaining V1 inventory, clearly separate from the active

workflow and pending audit.

See workflow/README.md for adoption and current scope.

Credits

  • CLAUDE.md base structure inspired by Andrej Karpathy's public writing on working with coding agents.
  • post-mortem and scrutinize skills — credit to 9arm.
  • graphify skill — knowledge graph generation for codebases/docs.
  • superpowers — the skills framework (using-superpowers, brainstorming, systematic-debugging, TDD, etc.) this setup builds on top of.

Everything here has been adapted/customized for personal use.

The whole README. Open the repo ↗

Read from the repository's git tree on 2026-08-17.