$npx -y skills add lucasfcosta/backpressured --skill goalRuns a backpressured automated development loop. Use it ONLY when the user explicitly mentions "backpressured" or "backpressure" — e.g. asking for a "backpressured" loop/run, "do this backpressured", or invoking it by that name — or runs the /backpressured:goal command. A bare
| 1 | # Backpressured |
| 2 | |
| 3 | ## Overview |
| 4 | |
| 5 | **You are the producer's backpressure of LAST resort — your job is to replace yourself with machines that say "no" first.** A human reading your code should be the last line of defense, never the first. Every "no" a machine can produce — a failing test, a type error, a lint rule, a benchmark regression, a reviewer's objection — must be produced *by the machine*, before you hand anything back. |
| 6 | |
| 7 | So you do not stop when the code "works." You drive the goal through the full loop below and stop **only** when every acceptance criterion and every quality check passes — or when you are genuinely blocked and can name exactly what is blocking you. |
| 8 | |
| 9 | ## When to Use |
| 10 | |
| 11 | - The user explicitly mentions "backpressured" or "backpressure" — asking for a backpressured loop/run, "do this backpressured", or invoking it by that name (brand-name mentions count too). |
| 12 | - The user runs the `/backpressured:goal` command. |
| 13 | |
| 14 | **Not for:** generic "implement X", "fix this bug", "build X and ship it", "open a PR", or a bare `/goal` request that does *not* explicitly mention backpressure. A finished, landed result being expected is not by itself enough — wait for one of the two signals above. Also not for one-line factual answers, pure research/exploration, or when the human wants to drive step by step. |
| 15 | |
| 16 | ## First: pin down "done" |
| 17 | |
| 18 | "Done" is defined by criteria, so make them explicit before anything else: |
| 19 | - **Acceptance criteria** — what must be true for the goal to be met. If the user didn't list them, infer them and state them up front. |
| 20 | - **Quality checks** — lint, tests, new behavior covered, project verification scripts, benchmarks, review. These apply on top of acceptance criteria, every time. |
| 21 | - **Project customization** — check for an optional `BACKPRESSURE.md` at the project root and fold its instructions in now (see [Figuring out the checks](#figuring-out-the-checks)). |
| 22 | |
| 23 | ## The Loop |
| 24 | |
| 25 | ```dot |
| 26 | digraph backpressure_loop { |
| 27 | "Goal received" [shape=doublecircle]; |
| 28 | "Write lightweight plan (approach + architecture, not details)" [shape=box]; |
| 29 | "Independent reviewer approves the plan?" [shape=diamond]; |
| 30 | "Revise plan" [shape=box]; |
| 31 | "Write the next patch" [shape=box]; |
| 32 | "Run every applicable check (incl. independent code review)" [shape=box]; |
| 33 | "All checks green?" [shape=diamond]; |
| 34 | "Fix the failure" [shape=box]; |
| 35 | "Genuinely blocked?" [shape=diamond]; |
| 36 | "STOP: name the blocker, do NOT mark done" [shape=octagon, style=filled, fillcolor=red, fontcolor=white]; |
| 37 | "All acceptance + quality criteria met?" [shape=diamond]; |
| 38 | "Run for real: curl, browser, full suite, full benchmarks, holistic independent review" [shape=box]; |
| 39 | "Open the PR" [shape=box]; |
| 40 | "Monitor: CI, comments, conflicts" [shape=box]; |
| 41 | "Landed clean with nothing outstanding?" [shape=diamond]; |
| 42 | "Done" [shape=doublecircle]; |
| 43 | |
| 44 | "Goal received" -> "Write lightweight plan (approach + architecture, not details)"; |
| 45 | "Write lightweight plan (approach + architecture, not details)" -> "Independent reviewer approves the plan?"; |
| 46 | "Independent reviewer approves the plan?" -> "Revise plan" [label="no"]; |
| 47 | "Revise plan" -> "Independent reviewer approves the plan?"; |
| 48 | "Independent reviewer approves the plan?" -> "Write the next patch" [label="yes"]; |
| 49 | "Write the next patch" -> "Run every applicable check (incl. independent code review)"; |
| 50 | "Run every applicable check (incl. independent code review)" -> "All checks green?"; |
| 51 | "All checks green?" -> "Fix the failure" [label="no"]; |
| 52 | "Fix the failure" -> "Genuinely blocked?"; |
| 53 | "Genuinely blocked?" -> "STOP: name the blocker, do NOT mark done" [label="yes"]; |
| 54 | "Genuinely blocked?" -> "Run every applicable check (incl. independent code review)" [label="no"]; |
| 55 | "All checks green?" -> "All acceptance + quality criteria met?" [label="yes"]; |
| 56 | "All acceptance + quality criteria met?" -> "Write the next patch" [label="no — keep going, do NOT hand back"]; |
| 57 | "All acceptance + quality criteria met?" -> "Run for real: curl, browser, full suite, full benchmarks, holistic independent review" [label="yes"]; |
| 58 | "Run for real: curl, browser, full suite, full benchmarks, holistic independent review" -> "Open the PR"; |
| 59 | "Open the PR" -> "Monitor: CI, comments, conflicts"; |
| 60 | "Monitor: CI, comments, conflicts" -> "Landed clean with nothing outstanding?"; |
| 61 | "Landed clean with nothing outstanding?" -> "Write the next patch" [label="no"]; |
| 62 | "Landed clean with nothing outstanding?" -> "Done" [label="yes"]; |
| 63 | } |
| 64 | ``` |
| 65 | |
| 66 | ### Phase 1 — Plan, then get it reviewed (before any |