Skip to main content
Glama

run_compile

Compiles a .vmf natively with vbsp, vvis, and vrad, returns per-stage findings, and stops at the first failure to catch leaks before wasted compile time.

Instructions

Compiles a .vmf with vbsp, vvis and vrad natively, and returns each stage's findings rather than its raw output. Takes minutes to hours: vvis on a large map is the slow part, and -fast exists for iteration. Stops at the first stage that fails, because vvis and vrad results are meaningless after a leak. Run read_vmf_lint first -- it catches before the compile most of what makes one fail.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "vmf": "<vmf>"
    },
    "name": "run_compile"
  }
}

Input Schema

NameRequiredDescriptionDefault
hdrNovrad -both, for an HDR and an LDR lightmap set.
vmfYesPath to the .vmf, absolute or relative to the repo root.
cullNoHammer++ only, and ON by default whenever the ++ chain runs. Culls unreferenced planes, vertices, brushes and brush sides even when no limit is reached yet. Measured on ttt_traps: -20.5% PLANES, -12.8% VERTEXES, -10.5% file size, with FACES and TEXINFO unchanged -- which is what distinguishes a prune from a broken map. Pass false to compile without it; read_map_geometry before and after is the check. Passing true with toolchain 'stock' is refused rather than ignored.
fastNovvis -fast and vrad -fast: minutes instead of hours, dev quality only.
gameNoGame profile id. This fork supports only 'l4d2', which is also the default; any other id is refused rather than answered about the wrong game.
stagesNoStages to run, in order. vbsp alone is enough to check a map seals.
confirmNo
toolchainNoWhich compilers to drive. The default 'plusplus' is ficool2's Hammer++ rebuild: much faster vvis and flags the stock chain does not have. Falls back to 'stock' when it is not installed, and the result says which one ran. 'stock' forces the compilers that ship with the game, which is what a comparison between the two needs.plusplus
timeoutMinutesNo

Output Schema

NameRequiredDescriptionDefault
okYes
bspYes
vmfYes
cullYes
gameYes
leakedYes
stagesYes
bspBytesYes
bspExistsYes
pointfileYes
toolchainYes
toolchainNoteYes
toolchainRequestedYes

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.0.24-alpha

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations supply only a title, so the description carries the full burden and does it well: expected duration (minutes to hours), the slow stage (vvis on large maps), fail-fast semantics with the reason (downstream stages are meaningless after a leak), and the toolchain fallback that reports which compiler actually ran.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Four sentences, front-loaded with purpose then cost then failure behavior then prerequisite; every clause carries operational information and there is no restatement of the tool name or title.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a long-running, multi-stage compile with an output schema, the description covers cost, failure and toolchain behavior thoroughly. The gap is the confirm parameter — a mutation-confirmation flag never explained anywhere — and the undocumented timeoutMinutes, which matter for a job that can run up to ten hours.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 78%, so the schema already documents hdr, cull, fast, game and stages in detail. The description reinforces only fast and stages, and adds nothing about confirm or timeoutMinutes, so it sits at the baseline for a well-described schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb and resource — compiles a .vmf through vbsp, vvis and vrad — and states the return shape ('each stage's findings rather than its raw output'), which differentiates it from read_compile_log, whose name implies log retrieval.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It explicitly names the prerequisite ('Run read_vmf_lint first'), justifies it (lint catches most compile failures), and gives a scoping rule for the stages parameter ('vbsp alone is enough to check a map seals') plus the -fast iteration tradeoff. When-to-use and when-to-narrow are both covered.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.