agentsclimarketplace

Turborepo foundation

Skill ahtishamshahzad/agent_dev_flow/.ai/skills/devops/turborepo-foundation

Use to plan a Turborepo setup after it's chosen — task pipeline definition (dependsOn), input/output declarations for correct caching, affected-only execution, and CI integration. Plans configuration; does not scaffold apps.From its SKILL.md

Install
npx -y skills add ahtishamshahzad/agent_dev_flow --skill turborepo-foundation

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 24 days oldThe repository was created 24 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.6 KB, 952 tokens by cl100k_base, as published. Nobody here has run it

Turborepo Foundation

Purpose

Configure Turborepo so its caching and orchestration are correct and fast: a task pipeline with accurate dependencies, precise input/output declarations, and affected-only execution wired into CI. Chosen upstream by monorepo-selection.

When to Use

  • After monorepo-selection chose Turborepo (over a relational package manager on top of pnpm/npm workspaces).
  • Not for workspaces-only monorepos or single repos.

Inputs

  • The workspace setup (pnpm-workspaces/npm-workspaces) and package task graph.
  • Build/test/lint task definitions per package.

Discovery Questions

  • What tasks exist per package (build, test, lint, typecheck) and how do they depend across packages?
  • What are each task's real inputs (source globs, config, env) and outputs (dist dirs) — the basis for correct caching?
  • Does CI want remote caching, and is it shared across the team/CI safely?

Responsibilities

  • Define the task pipeline: dependsOn relationships (^build for upstream builds), so tasks run in dependency order.
  • Declare inputs and outputs precisely per task — this is where caching correctness lives: missing inputs → stale cache serving wrong results; missing outputs → cache misses. Include config files and relevant env in inputs.
  • Enable affected-only execution so CI runs only what changed and its dependents (ci-cd), and local runs reuse cache.
  • Configure caching: local by default; remote cache only with a deliberate, secured setup (tokens as secrets — secrets-management) and awareness that a poisoned/misconfigured cache is a correctness and supply-chain risk.
  • Keep task scripts in the packages; Turborepo orchestrates, it doesn't own the logic.

Required Workflow

  1. Enumerate tasks + cross-package dependencies.
  2. Define the pipeline with dependsOn.
  3. Declare accurate inputs/outputs per task (verify cache correctness).
  4. Wire affected-only execution into CI.
  5. Configure caching (local; remote only if secured); validate no stale-cache correctness bugs.

Decision Rules

  • Caching correctness depends on complete input/output declarations — under-declaring inputs is a silent-wrong-result bug, over-declaring just misses cache.
  • Remote caching is opt-in and secured; treat cache artifacts as trusted-build inputs (supply-chain surface — ../../security/dependency-security).
  • Orchestrate; don't relocate build logic into the pipeline config.
  • Validate cache hits produce identical results to clean builds before trusting them in CI.

Rules

  • Cache tokens/secrets via the secret store, never committed (secrets-management).
  • Pipeline changes reviewed — cache correctness is easy to break subtly.
  • Turborepo config stays proportional; no orchestration of tasks that don't exist.

Anti-Patterns

  • Under-declared inputs serving stale cache as fresh results.
  • Committed remote-cache tokens.
  • Build logic migrated into turbo.json instead of package scripts.
  • Trusting remote cache without validating correctness.
  • Configuring pipelines for tasks/packages that aren't there.

Validation Checklist

  • Task pipeline with correct dependsOn.
  • Accurate inputs/outputs per task (cache-correct).
  • Affected-only execution wired to CI.
  • Caching configured; remote (if any) secured; correctness validated.
  • Secrets handled via the store.

Definition of Done

A Turborepo pipeline with correct task dependencies and precise, validated input/output declarations, affected-only CI execution, and secured caching — orchestrating existing package tasks without owning their logic.

Related Skills

monorepo-selection, pnpm-workspaces, npm-workspaces, ci-cd, github-actions, secrets-management, ../../security/dependency-security.

Related Knowledge

../../../knowledge/ (task graph, CI needs).

Related References

../../../references/devops/ (pipeline patterns, when populated).

Context Loading Guidance

  • Requires: workspace setup, per-package task graph with inputs/outputs.
  • Does not require: app feature code, unrelated infra.
  • May load: ci-cd, secrets-management (remote cache).
  • Stop when: the pipeline + caching + affected-only CI are configured and validated.

Token Efficiency Guidance

The task × (dependsOn, inputs, outputs) table is the artifact; get those right and the config follows.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,861. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.