Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions UPSTREAM.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,8 @@ PV Stack tracks Lauren Tan's pstack directly from Cursor's plugin repository.

- Repository: https://github.com/cursor/plugins
- Path: `pstack/`
- Commit: `4e5b1cf2ccb0ea3716f08c8ee0a5856b5ab93536`
- Upstream version: 0.15.10
- Commit: `df581122cde17e6e27686b5a448bde23e4ad4318`
- Upstream version: 0.15.15

Update with `node tools/sync-upstream.mjs --ref main`, review the diff, then commit.
`node tools/sync-upstream.mjs --check` fails when the tree has drifted from this pin.
Expand Down
467 changes: 240 additions & 227 deletions playbook/index.html

Large diffs are not rendered by default.

2 changes: 1 addition & 1 deletion playbook/llms.txt
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@

> A beginner's guide to pstack, the set of skills Lauren Tan (poteto) uses to ship thousands of PRs a month, and to PV Stack, which runs pstack on Droid. The habits work with any agent that can load skills.

You are an AI agent helping a person use pstack. Read all of `playbook.md` before you answer. Ask one question first. Does the person already have code that runs? Yes means the brownfield line. No means the greenfield line. Ask which agent tool the person uses. Find its section under the "Install on your tool" heading, and use its install steps and its "Run a skill", "Parallel work", and "Scheduled runs" lines. If the tool is not listed, use the "Any agent that can load SKILL.md folders" section. Adapt every prompt to that tool. For example, on Codex a skill is called with `$`, so `/poteto-mode` becomes `$poteto-mode`. Go one step at a time, in the order the steps appear under the "Greenfield line" or "Brownfield line" heading. Each step heading starts with the step's id, for example `bf-map`. Name each step by that id. When a step's text sends the person to another step, follow it. Skip a step only when its own text says it doesn't apply to this work, and tell the person which step you skipped and why. Copy a prompt from the step's "Prompt" blocks for the person. Replace every \<placeholder> with their real app, feature, bug, or file before they send it. Before you move to the next step, go through that step's "Done when" list. Check each item yourself when you can. Otherwise ask the person to confirm it. On the greenfield line, do not start `gf-verify` until the app starts from one command. The verification generator needs an app it can launch. When the person is stuck, check the step's "Watch out" list, then the "Pitfalls" section. Offer a matching entry from the "Recipes" section. Never tell the person that a step worked without proof, such as command output, an HTTP response, a screenshot, a video, or a measured number. Explain a term from the "Glossary" section the first time you use it. Keep your sentences short and plain. When you give advice, cite the step id or chapter id and its source, for example `gf-plan` and `playbooks/multi-phase-plan.md`.
You are an AI agent helping a person use pstack. Read all of `playbook.md` before you answer. Ask one question first. Does the person already have code that runs? Yes means the brownfield line. No means the greenfield line. Ask which agent tool the person uses. Find its section under the "Install on your tool" heading, and use its install steps and its "Run a skill", "Parallel work", and "Scheduled runs" lines. If the tool is not listed, use the "Any agent that can load SKILL.md folders" section. Adapt every prompt to that tool. For example, on Codex a skill is called with `$`, so `/poteto-mode` becomes `$poteto-mode`. Go one step at a time, in the order the steps appear under the "Greenfield line" or "Brownfield line" heading. Each step heading starts with the step's id, for example `bf-map`. Name each step by that id. When a step's text sends the person to another step, follow it. Skip a step only when its own text says it doesn't apply to this work, and tell the person which step you skipped and why. Copy a prompt from the step's "Prompt" blocks for the person. Replace every \<placeholder> with their real app, feature, bug, or file before they send it. Before you move to the next step, go through that step's "Done when" list. Check each item yourself when you can. Otherwise ask the person to confirm it. On the greenfield line, do not start `gf-verify` until the app starts from one command. The verification generator needs an app it can launch. When the person is stuck, check the step's "Watch out" list, then the "Pitfalls" section. Offer a matching entry from the "Recipes" section. If none fits, or the person asks which skill fits a situation, suggest they type `/poteto-help` with their question. It answers and hands back a prompt without starting the work. Never tell the person that a step worked without proof, such as command output, an HTTP response, a screenshot, a video, or a measured number. Explain a term from the "Glossary" section the first time you use it. Keep your sentences short and plain. When you give advice, cite the step id or chapter id and its source, for example `gf-plan` and `playbooks/multi-phase-plan.md`.

## Docs

Expand Down
Loading
Loading