agentleFS
vosslabvosslab/vosslab-skills/skills/management/html-game-parallel-builder/SKILL.md

html-game-parallel-builder

Skill5 starsChanged 21 days agoMITView on GitHub

Use when the user requests a parallel multi-agent build or major workstream for a modular TypeScript browser game, including integrated serving, smoke testing, export, or Pages delivery.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 7 are distinctive, and 10 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use for a time-pressured, modular TypeScript browser-game build where multiple agents must author interacting modules and the live local …

    line 8
  2. Very specific:

    Before contracts or dispatch, invoke game-design-expert to establish the shared design brief: intended experience, core actions and …

    line 16
  3. Very specific:

    Read references/BATCH_DISPATCH.md before file assignment.

    line 30
  4. Very specific:

    Default to a GitHub Pages-ready build in dist/.

    line 38
  5. Very specific:

    Run Batch 1 sequentially, then Batches 2-4 in the ownership order from references/BATCH_DISPATCH.md.

    line 56
  6. Very specific:

    After every batch, run the type-check and Playwright smoke described in references/SMOKE_AND_GOTCHAS.md.

    line 58
  7. Very specific:

    After the final batch, run the full playthrough and use the proportional integrated-review and fix loop in references/STEP_DETAILS.md until …

    line 62
  8. Distinctive:

    Never dispatch coding agents before contracts exist.

    line 68
  9. Distinctive:

    Never define contracts or assignments before the shared design brief exists.

    line 69
  10. Very specific:

    Never declare a batch green without the exact successful tsc --noEmit evidence and its required smoke result.

    line 70
  11. Distinctive:

    Never skip a between-batch smoke or proceed past its failure.

    line 72
  12. Very specific:

    Never locally redeclare a cross-module shape.

    line 73
  13. Distinctive:

    Keep coding ownership exclusive.

    line 75
  14. Distinctive:

    Keep every milestone, prerequisite, assessment, correction, and final acceptance executable by the manager and subagents without waiting …

    line 77
  15. Distinctive:

    Do not blend build identities, weaken type safety with casts, or replace browser evidence with a claim that the game is green.

    line 82
  16. Very specific:

    Read references/QUALITY_GUARDRAILS.md for the complete red flags and rationale.

    line 85
  17. Very specific:

    Read the target repository's docs/FUN_VIBES_DESIGN_STYLE.md and docs/PLAYFUL_TRAINING_GAME_STYLE.md when present.

    line 95

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Web Game Parallel Build

Use for a time-pressured, modular TypeScript browser-game build where multiple agents must author interacting modules and the live local preview is the immediate target. It is not for a simple edit, server application, heavyweight multi-page build, or work where one agent can finish promptly. Parallelism cuts wall time only when contracts and integration gates remain intact.

Required routing

Before contracts or dispatch, invoke game-design-expert to establish the shared design brief: intended experience, core actions and decisions, surprise and progression, visual identity and visible consequence, cross-module dependencies, and evidence. Keep that concise brief in the project's existing planning documentation. It is the creative source for contracts, assignments, and integrated review.

Also invoke delegate-manager-to-subagents for ownership and evidence discipline, parallel-plan for lane sizing and ownership splits, and typescript-engineer for all type design. Read the TypeScript skill's game type, modular-type, strict-mode, and opaque-type references before defining cross-module contracts. This skill owns browser-game orchestration, not TypeScript design rules.

Read references/BATCH_DISPATCH.md before file assignment. It owns the A0-E layout, batch-specific prompt requirements, exclusive file ownership, and the type-design pause-and-route procedure. Use templates/agent_prompt_template.md for every coding assignment.

Build and release identity

Default to a GitHub Pages-ready build in dist/. Use a portable single-file export in output_single_file/ only when explicitly requested. Keep the three paths separate: run_web_server.sh serves the live preview, build_github_pages.sh produces the Pages build, and export_single_file.sh produces the optional portable artifact. Read references/BUILD_ARTIFACTS.md for ownership, starter-template parity, copied artifacts, and output constraints. Read references/GITHUB_PAGES_DEPLOY.md only when configuring or diagnosing deployment.

Control loop

  1. Establish the shared design brief without waiting for human input, define a first playable slice, and write real shared contracts in src/types/ before any coding agent starts. The slice orders work; it does not reduce the requested game to a smaller final outcome.
  2. Configure the supported scaffold and establish a clean npx tsc --noEmit -p tsconfig.json gate.
  3. Run Batch 1 sequentially, then Batches 2-4 in the ownership order from references/BATCH_DISPATCH.md.
  4. After every batch, run the type-check and Playwright smoke described in references/SMOKE_AND_GOTCHAS.md. A failing check stops progression: repair the responsible module or contract, then rebuild and re-test.
  5. After the final batch, run the full playthrough and use the proportional integrated-review and fix loop in references/STEP_DETAILS.md until clean.

Load-bearing safeguards

  • Never dispatch coding agents before contracts exist.
  • Never define contracts or assignments before the shared design brief exists.
  • Never declare a batch green without the exact successful tsc --noEmit evidence and its required smoke result.
  • Never skip a between-batch smoke or proceed past its failure.
  • Never locally redeclare a cross-module shape; route it to the contract owner and typescript-engineer.
  • Keep coding ownership exclusive. The manager writes no game code except the documented Batch 1 foundation boundary.
  • Keep every milestone, prerequisite, assessment, correction, and final acceptance executable by the manager and subagents without waiting for a human. Scale independent assessment to the change's importance and uncertainty; do not add routine artwork bake-offs or extra reviewers for small, settled fixes.
  • Do not blend build identities, weaken type safety with casts, or replace browser evidence with a claim that the game is green.

Read references/QUALITY_GUARDRAILS.md for the complete red flags and rationale; read references/SMOKE_AND_GOTCHAS.md when web platform behavior fails. Use references/STEP_DETAILS.md for conditional setup, scope, and repair detail. Definitions, naming, and capacity route respectively through references/DEFINITIONS.md, references/NAMING_GUARDRAILS.md, references/CAPACITY_AND_SIZING.md.

Read the target repository's docs/FUN_VIBES_DESIGN_STYLE.md and docs/PLAYFUL_TRAINING_GAME_STYLE.md when present. Their canonical sources live in starter-repo-template/templates/typescript/docs/; this skill bundles neither guide. When absent, use game-design-expert and the shared brief.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.