agentleFS
AshExplainedAshExplained/roblox-skills/roblox-triage/SKILL.md

roblox-triage

Skill5 starsChanged 5 months agoMITView on GitHub

Triage Roblox game slices and bug reports through a small state machine so only fully-specified, playtestable work gets built. Use when the user wants to review incoming slices or bugs, decide what to build next, prepare a slice for building, mark work ready, or manage the roadmap/issue workflow. Keeps Claude from picking up vague, half-specified work and building the wrong thing.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every triaged item carries exactly one category and one state role.

    line 31
  2. Very specific:

    Gather context: Read the item fully (and prior triage notes so you don't re-ask).

    line 49
  3. Very specific:

    Reproduce (bugs only): Before anything else, try to reproduce in a Studio playtest and read get_console_output.

    line 51
  4. Very specific:

    Sharpen (if needed): If the item is vague, run a short grilling pass (see roblox-game-ideas grilling mode) to pin down the decision.

    line 52
  5. Very specific:

    If the user says "mark X ready to build", trust them and apply it after checking for conflicting state roles and the UI design gate.

    line 61

≈ “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

Roblox Triage

Move slices and bug reports through a small state machine so that anything marked "ready to build" is actually specified enough to build in one focused pass — and playtestable when done.

The tracker location and label vocabulary should have been provided to you in CLAUDE.md / docs/agents/. If not, run setup-roblox-skills. If the tracker is GitHub, prefix any AI-written comment with > *Generated by AI during triage.*.

Reference

  • BUILD-BRIEF.md — how to write a durable build brief for a ready-to-build slice.

Roles

Two category roles:

  • bug — something is broken in the place.
  • enhancement — a new slice / capability.

Five state roles:

  • needs-triage — not yet evaluated.
  • needs-info — waiting on the user for a decision or detail.
  • ready-to-build — fully specified; an agent can build it in one pass with no missing context.
  • needs-human — needs the user (a Studio action, a design call, a balance judgment, a purchase test).
  • wontfix — will not be actioned.

Every triaged item carries exactly one category and one state role. These are canonical names; the actual label/status strings are in docs/agents/triage-labels.md.

Normal flow: needs-triage → (needs-info ↔ needs-triage) → ready-to-build / needs-human / wontfix. The user can override; flag anything unusual and ask first.

UI design gate

A slice that changes user-facing UI cannot become ready-to-build until an approved Roblox UI design artifact exists for it (see roblox-game-ux for what that artifact contains: layout, states, and mobile/tablet/desktop behavior, marked Status: approved). If it is missing or still draft, the slice becomes needs-human with a note that it needs the UI design step first.

Invocation

The user invokes this in natural language: "what should we build next?", "triage this bug", "is this slice ready?", "mark the coin-pickup slice ready to build". Interpret and act.

Show what needs attention

Present three buckets, oldest first: untriaged, needs-triage, and needs-info with new user input. Show a one-line summary each and let the user pick.

Triage one item

  1. Gather context. Read the item fully (and prior triage notes so you don't re-ask). Connect via roblox-studio-workflow and inspect the relevant scripts/instances. For UI items, check the design artifact.
  2. Recommend. State your category + state recommendation with reasoning and a short summary of the relevant place state. Wait for direction.
  3. Reproduce (bugs only). Before anything else, try to reproduce in a Studio playtest and read get_console_output. A confirmed repro (with the failing code path) makes a far stronger build brief; no repro is a strong needs-info signal. Hand hard bugs to roblox-debugging-bugfix.
  4. Sharpen (if needed). If the item is vague, run a short grilling pass (see roblox-game-ideas grilling mode) to pin down the decision.
  5. Apply the outcome:
    • ready-to-build — post a build brief (BUILD-BRIEF.md). For UI, include the approved artifact path.
    • needs-human — same structure, but say why it can't be delegated.
    • needs-info — post triage notes: what's established, and the specific questions still open.
    • wontfix — brief reason; for enhancements, note why so it isn't re-suggested.

Quick override

If the user says "mark X ready to build", trust them and apply it after checking for conflicting state roles and the UI design gate. If a UI artifact is missing, say exactly what's missing and ask for explicit override before applying ready-to-build.

Next

Build the ready-to-build slice with its phase specialist (roblox-luau-architecture, roblox-combat-systems, roblox-ui-implementation, roblox-datastore-persistence, etc.), then verify with roblox-playtest-qa and roblox-mobile-playtest. If a bug was triaged, hand it to roblox-debugging-bugfix. Return to roblox-game-development-lifecycle to pick the next slice.

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.