Lauren Tan's pstack is an opinionated Cursor skill stack that improves agent outcomes. Michael Denyer's pstack-claude ports it to Claude Code, Codex, Pi, GitHub Copilot, Muse, and other agent harnesses. This fork adds what the port had to drop: a role can run on a model from another provider. It ships Claude Code and Codex only and drops pstack-claude's Pi, GitHub Copilot, and Muse builds.
Tell poteto-mode your goal and it will invoke the correct workflow for the task. It keeps your code concise, simple and verified. When a role's model belongs to another provider, the run-role skill launches it as a supervised Orca worker through the bundled orca-delegate skill; roles on the host provider stay native subagents.
For concurrency bugs and invariants that tests cannot reach, /create-formal-verification-skill picks the riskiest property in your code, verifies it with a tool that fits, and leaves a project skill the next agent reruns. The separate agent-formal-verify plugin goes deeper on the method, with guided TLA+ model checking, Lean proofs, and Kani for Rust.
Claude Code
Run in Claude Code 2.1.292 or later:
/plugin marketplace add dev-dingguri/pstack-orca
/plugin install pstack@pstack-orca
Codex
Run in your terminal:
codex plugin marketplace add dev-dingguri/pstack-orca
codex plugin add pstack@pstack-orcaOther harnesses
OpenCode (the CLI and the desktop app), Gemini CLI, Prime Agent, and skills-only Codex installs read skills from ~/.agents/skills/. Run in your terminal:
git clone https://github.com/dev-dingguri/pstack-orca
cd pstack-orca
mkdir -p ~/.agents/skills
for s in plugins/pstack/skills/*/; do
target=~/.agents/skills/"$(basename "$s")"
test -e "$target" || test -L "$target" || ln -s "$(pwd)/$s" "$target"
doneTo update, run git pull in the clone, then run the for loop again to link skills added since the last run. These installs have no routing hook, so ask for poteto-mode by name or add a standing instruction to the harness's instruction file. See shared installation for per-harness notes and uninstalling.
pstack routes tasks to poteto-mode automatically. At session start, a hook injects an instruction that points the agent at it: the plugin's hook on Claude Code and Codex. Codex runs the hook only after you trust it through /hooks. Other harnesses have no hook.
Run setup-pstack to turn automatic routing on or off, change model defaults, or set a reasoning effort per role (arena runners: opus@claude @xhigh, gpt-6-astra@codex @high). Every model entry is model@provider; the default panel mixes Claude and Codex models, and single-model roles stay on the host. Cross-provider entries need Orca installed; a start that fails is reported and waits for your decision. In Claude Code, use /pstack:setup-pstack.
The plugin keeps the pstack name and the /pstack:* skill names. Uninstall pstack@pstack-claude first; two installs would both register the same skills and hook.
Use poteto-mode to fix the search filter resetting when I change pages.
For a bug, it reproduces the failure, uses how and why to investigate, delegates the fix, then reruns the failing case. If the fix crosses a function boundary, it brings in architect before implementation. You receive the fix and the failing and passing evidence.
Other playbooks cover planning, features, refactoring, performance issues, investigations, prototypes, PR maintenance, shipping, and longer projects.
- Multi-provider roles
- Skills and slash commands
- Runtime setup
- Models and dependencies
- Maintenance and port scope
pstack has no server or telemetry. Anything its skills ask your agent to read, including session transcripts, goes to your model provider. Scripts run locally, and PR tools use your GitHub CLI login.
Thanks for helping make this fork better. Bug reports, documentation fixes, and runtime improvements are welcome. Workflow changes belong upstream in pstack-claude or pstack; provider routing belongs here. See CONTRIBUTING.md for the checks and where your change belongs. Report vulnerabilities privately as described in SECURITY.md.
This fork, including its modifications and additions, is also MIT-licensed. Original pstack © 2026 Lauren Tan; pstack-claude port © 2026 Michael Denyer; imported cursor-team-kit and thermos skills © 2026 Cursor. See LICENSE-cursor-team-kit, LICENSE-thermos, and NOTICE.md.
