pstack for WorkBuddy
pstack is poteto's agent style packaged as skills: concise, detailed replies, deliberate subagents, unslopped prose, simple code, and verified work. /poteto-mode matches a task to a playbook, copies the playbook's steps into the todo list, and runs the other skills as the steps need them. This bundle is the community WorkBuddy port of pstack 0.15.32.
What is installed
All 58 pstack skills sit next to this file in skills/<name>/SKILL.md. WorkBuddy discovers them as their own skills, so each one is a slash command under its bare name, such as /poteto-mode. This bundle is not the plugin, so there is no pstack: prefix.
The skill text uses the plugin form /pstack:<name>. When a skill says /pstack:<name>, use /<name>. When it says pstack:poteto-agent or pstack:comment-sicko, see "Missing compared with the full plugin" below.
Start
/setup-pstack: pick which WorkBuddy models pstack uses per role and the reasoning effort. It writes~/.workbuddy/pstack-models.md. You can skip it. Subagents then inherit the current model./poteto-mode <task>: the main entry point./poteto-help <question>: guidance on setup and on picking a skill, playbook, or principle.
If a command is not in the list, read skills/<name>/SKILL.md in this folder and follow it. Do the same when one skill tells you to run another.
Paths
This folder is the <plugin> root that the skills mean. It holds skills/, docs/, agents/, LICENSE.md, and NOTICE.md. Read docs/harness/workbuddy.md before you follow a skill that asks for it. It maps Cursor tools, subagents, models, and paths to WorkBuddy. The user guide is docs/guide/README.md.
Skills
Start here
/setup-pstack: Configure which WorkBuddy models pstack uses per role and at what reasoning effort./poteto-mode: poteto's agent style for concise, detailed responses, deliberate subagents, unslopped prose, simple code, and verified work./poteto-help: Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task.
Principles (one-rule skills poteto-mode cites; you rarely invoke them directly)
/principle-attack-the-premise: Write the premise down and test it directly before you write another fix that assumes it./principle-boundary-discipline: Concentrate guards at system boundaries (CLI, config, network, external APIs); trust internal types and keep business logic in pure functions./principle-build-the-lever: Build the tool that does it or proves it (codemod, script, generator, or a skill your subagents follow) instead of working by hand./principle-encode-lessons-in-structure: Encode the rule as a lint, metadata flag, runtime check, or script instead of more text./principle-exhaust-the-design-space: Build 2-3 competing prototypes and compare side by side before committing./principle-experience-first: Choose user delight over implementation convenience; ship fewer polished features over more rough ones./principle-explain-the-number: Find what limits it, and rule out that it measured something other than the work you think./principle-fix-root-causes: Trace each symptom back to the first place the system went wrong and fix it there./principle-foundational-thinking: Get the data structures right so downstream code becomes obvious./principle-guard-the-context-window: Route bulk to subagents; keep summaries in the main thread, not raw payloads./principle-laziness-protocol: Bias toward deletion and the smallest change that solves the problem./principle-make-operations-idempotent: Converge to the same end state regardless of partial prior runs./principle-migrate-callers-then-delete-legacy-apis: Migrate callers and delete the old API in the same wave instead of preserving compatibility layers./principle-minimize-reader-load: Count layers between question and answer, and hidden state in the reader's head; collapse one-caller wrappers and shrink mutable scope./principle-model-the-domain: Apply when writing stateful logic, or when code branches a lot or repeats a shape assumption across files./principle-never-block-on-the-human: Apply when tempted to ask 'should I do X?' on reversible work./principle-outcome-oriented-execution: Converge on the target architecture; don't preserve smooth intermediate states with throwaway compatibility code./principle-prove-it-works: Verify against the real artifact (run the feature, read the actual value, inspect the diff), not a proxy, self-report, or 'it compiles.'/principle-redesign-from-first-principles: Apply when integrating a new requirement into an existing design./principle-separate-before-serializing-shared-state: Eliminate the sharing first; serialize structurally only when one shared writer is a real invariant./principle-sequence-verifiable-units: Break work into small units that each end in a verifiable state, check each before the next, and order delivery so the sequence proves itself to a reviewer./principle-subtract-before-you-add: Remove dead code, redundant validators, and stub references first, then build on the simpler base./principle-test-behavior-not-implementation: Assert what the code's users can observe, so that the test fails when that behavior breaks and keeps passing when only the internals change./principle-type-system-discipline: Make illegal states unrepresentable, brand semantic primitives, parse external data at boundaries, exhaust variants, derive from authoritative schemas.
Workflow skills
/architect: Sketch types, signatures, and module structure before code, then stay in the loop while implementation fills in./arena: Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it./automate-me: Drafts or revises a personal -mode skill via create-skill + unslop, optionally pulling fresh evidence from recent transcripts./benchmark-checklist: Vet a perf measurement (limiter, tuning, limits, errors, repeatability, relevance, and whether the work happened) before you report or act on it./blast-radius: Find what a change could break somewhere else before it ships, beyond the diff, and prove the one fact it's safe because of by running real code instead of writing it up./bro: Restate the last message in plain human language, with no jargon./correct: Find the mistakes agents keep repeating in this repo and make each one impossible./create-formal-verification-skill: Generate a project-local skill that formally verifies the code where a bug costs the most and tests reach the least, in any language, with whichever verification tool.../create-verification-skill: Generate a project-local verification skill that drives your app the way a user does — any language, framework, or platform./figure-it-out: Design an auditable playbook when no narrower one fits: a large migration, an ambitious multi-part change, or work a human reviews after stepping away./how: Explains subsystem architecture, runtime flow, onboarding mental models./in-your-own-words: Restate in your own words what you think the user's goals are and what problem they're trying to solve./interrogate: Multiple LLM reviewers challenge changes from independent angles./ketchup: Catch the user up on what changed in this chat since their last message, in plain words and small visuals./ketchup-video: Catch the user up like /ketchup, with a short silent video of slides for what happened and the waiting calls as text under it./maintain-formal-verification-skill: Bring a project's formal verification back in line with the code after the code changes: find the commits that touched verified code, update the model, harness, or pro.../maintain-verification-skill: Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most on.../make-bot-ui: Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing.../no-comments: Spawn Comment Sicko, fix accepted findings, and offer encodings for claimed constraints./recall: Reconstruct your recent working context from your own chat history, live state, and the shared record (user reports, prior fixes, incidents), then hand back a tight cu.../reflect: Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill./sharpen: Make a task deeper, not bigger, then rewrite the prompt./show-me-your-work: Keep a reviewable decision trail for long-running or unattended work: a TSV log with one row per decision (what, why, evidence, result)./statechart: Draw how any system moves between states, such as a product flow, a protocol, or a service, as a Mermaid statechart from a TLA+ model./swarm: Fan out N parallel workers, drain them, and return one report./tdd: Use only when the user explicitly asks for TDD, a failing test, or a regression test, OR when the bug has an obvious cheap local test target./teach: Explain a body of work plainly so a person actually understands it./technical-writing: Layered technical-writing standard: Diátaxis structure, Google developer style sentences, STE instruction rules, Global English syntax./typescript-best-practices: TypeScript best practices./unslop: Cut AI tells from any writing./why: Discovers available MCPs and queries each evidence category (source control, issue tracker, long-form docs, real-time chat, infrastructure observability, error trackin...
Missing compared with the full plugin
- The plugin's hooks are not loaded from a skill install.
/poteto-modedoes not stay on across turns, so invoke it again on each new task. TheSessionStarthook that injects~/.workbuddy/pstack-models.mdis also absent. Read that file yourself when a skill needs a model role. - The plugin agents
poteto-agentandcomment-sickoare not registered. Their definitions are inagents/as reference docs. When a skill asks forsubagent_type: "pstack:poteto-agent"or"pstack:comment-sicko", spawn ageneral-purposeAgent. Paste the body ofagents/poteto-agent.mdoragents/comment-sicko.mdinto its prompt, along with the absolute path ofskills/poteto-mode/SKILL.mdin this folder. Say that you substituted the agent.
For sticky mode and the real agents, install the plugin. Add the marketplace pstack-pro/pstack-workbuddy as a git repo, then install pstack@pstack-workbuddy:
/plugin marketplace add pstack-pro/pstack-workbuddy
/plugin install pstack@pstack-workbuddy
SkillHub packaging notes
SkillHub rejects files without an extension, so this bundle renames three files: LICENSE is LICENSE.md, NOTICE is NOTICE.md, and the watch-pr launcher skills/poteto-mode/scripts/watch-pr/watch-pr is watch-pr.ts. It also leaves out images, bun.lock (bun recreates it on install) and the statechart script tests, and ships skills/show-me-your-work/references/decision-log-template.tsv as decision-log-template.txt (same tab-separated content). When a skill says to run scripts/watch-pr/watch-pr, run bun skills/poteto-mode/scripts/watch-pr/watch-pr.ts with the same arguments.
Credits
pstack is by Lauren Tan (poteto), upstream at https://github.com/cursor/plugins/tree/main/pstack. The skill, agent, and guide text is Copyright (c) 2026 Lauren Tan and is used under the MIT license in LICENSE.md. The WorkBuddy port is by pstack-pro (https://github.com/pstack-pro/pstack-workbuddy). See NOTICE.md. This bundle is not affiliated with Cursor or Tencent.
