codometer-triage — скилл для Claude Code: что делает и как установить
Скилл codometer-triage разбирает упавший codometer --check: отличает превышение лимита от устаревшего отчёта и от прогона, который не смог завершиться, и говорит, что чинить.
MIT 363 установок обновлён 2026-10-24
npx skills add organizzolini/codebase --skill codometer-triage -a claude-codeЧто делает codometer-triage
Скилл включается, когда codometer --check вернул ненулевой код, и первым делом учит различать три находки: превышение лимита, устаревший закоммиченный отчёт и прогон, который не смог завершиться. Дальше для каждой свой ремонт.
Уровень решает, падает ли что-то вообще. warn печатает превышение и оставляет код выхода как был; fail даёт 1 и только при --check limits. Бюджет комментариев приходит обычной записью в limits[] против метрики custom.<label>, но в instances указаны файл и строка каждого блока. Считать стоит с оговоркой: сканер строк в shell, TOML, SQL и HCL примет маркер комментария внутри строкового литерала за настоящий. Детали про объявление входов и флаги — в соседних скиллах codometer-configure и codometer-measure.
Поднимать value лимита, чтобы чек позеленел, — не тот ремонт: вместе с уликой стирается обещание о форме кодовой базы. Сжатые размеры зависят от версии Node, поэтому отчёт, снятый на одной версии, на другой читается как устаревший без единой правки в коде. --output-json или --output-markdown вместе с --check reports отклоняют сразу.
Какие задачи решает
- Понять, почему
codometer --checkвернул ненулевой код - Разжать бюджет комментариев в разросшемся блоке
- Освежить закоммиченный отчёт вместо правки лимита
- Найти лимит, привязанный к метрике, которой прогон не считал
- Починить конфигурацию, отклонённую до замера
Примеры запросов агенту
codometer --checkупал на лимите размера бандла — разбери отчёт и скажи, что резать в коде, а не в конфиге.- Отчёт в репозитории числится устаревшим, хотя код не менялся. Проверь, на какой версии Node его снимали.
- В прогоне три записи в
failures, разложи их по причинам и предложи правки.
Как установить codometer-triage
1Через Skills CLI (Vercel)
npx skills add organizzolini/codebase --skill codometer-triage -a claude-code2Вручную: положить файлы
~/.claude/skills/codometer-triage/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/codometer-triage/
1Через Skills CLI (Vercel)
npx skills add organizzolini/codebase --skill codometer-triage -a codex2Вручную: положить файлы
.agents/skills/codometer-triage/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add organizzolini/codebase --skill codometer-triage -a cursor2Вручную: положить файлы
.cursor/skills/codometer-triage/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install organizzolini/codebase projects/ic-suite/codometer/codometer-agents/skills/codometer-triage2Вручную: положить файлы
.github/skills/codometer-triage/SKILL.md1Через Skills CLI (Vercel)
npx skills add organizzolini/codebase --skill codometer-triage -a gemini-cli2Командой платформы
gemini skills install https://github.com/organizzolini/codebase --path projects/ic-suite/codometer/codometer-agents/skills/codometer-triageВручную: положить файлы
.devin/skills/codometer-triage/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add organizzolini/codebase --skill codometer-triage -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Вопросы о codometer-triage
Как установить codometer-triage в Claude Code?
Выполните npx skills add organizzolini/codebase --skill codometer-triage -a claude-code. Файл скилла появится в ~/.claude/skills/codometer-triage/SKILL.md.
Что делает codometer-triage?
Помогает агенту прочитать отчёт codometer --check и выбрать ремонт под тип находки: превышение лимита, устаревший отчёт или незавершённый прогон. Отдельно разбирает бюджет комментариев и ошибки конфигурации.
Чем превышение лимита отличается от устаревшего отчёта?
Это разные находки. Превышение означает, что метрика ушла за объявленный лимит. Устаревание — расхождение закоммиченного отчёта со свежим замером, и лимит там может быть ни при чём.
Исходный текст
SKILL.md · MIT
--- name: codometer-triage description: Act on a codometer limit breach, a comment budget breach, a stale committed report, or a run that failed outright. Use when a codometer --check exits non-zero, when a size or count limit was breached, when a comment block ran past its word, line, or character budget, when a committed report reads stale against a fresh measurement even though nothing changed, when a limit fails to bind or an input matches no files, or before reaching for a repository's limit value to make a failing check pass. license: MIT --- # Acting on a codometer failure A `--check` run reports three different kinds of finding, and they call for three different responses. Reading which one occurred, rather than reacting to "the check failed," is most of this skill. ## A breach A metric went past a limit declared for it. The report names the metric, the limit's value, and its severity: - **`warn`** prints the breach and leaves the exit code alone. Nothing about the run fails; it is advance notice that a number is moving in a direction worth watching. - **`fail`** exits `1`, and only when the run asked `--check limits` for a gate at all — a `codometer measure --config <path>` measuring the same breach reports nothing wrong, because nothing asked it to gate. **The instinctive fix — raising the limit's `value` — is not an option.** The limit exists to keep a promise about the codebase's shape; loosening it on the same change that broke the promise erases the evidence the check existed to produce, and the next breach starts from a codebase that is already worse. Reduce what is actually being measured instead: - A size breach: shrink the input's own output — trim a dependency, split a bundle, delete dead code — rather than the number that reports it. The `codometer-configure` skill covers declaring an input and pointing it at compiled output correctly, which is a common reason a size looks larger than the code that produced it: an input measuring more than it should. - A convention-count breach (too many of something, or too few): the count is usually accurate — the fix is in the code the counter is watching, not in the counter's `value`. - A **comment budget** breach is a custom statistic like any other, gated by an ordinary `limits[]` entry against its `custom.<label>` metric — but its `instances` name a file and a line for each breaching block rather than only a total, because the thing measured is one block rather than a repository-wide count. Condense that block, or move the detail into documentation that has room for it. Check which reader measured it before trusting the count — shell, TOML, SQL, and HCL use a line scanner that reads a comment marker inside a string literal as a comment, and Python's comments go unmeasured entirely when the interpreter cannot be run. The `codometer-configure` skill has both caveats in full. - **A comment budget a language is not covered by is not a breach at all** — it is simply not measured. If a repository added a language codometer can parse comments in but nobody added a `comment` selector for it, that language's blocks run with no budget until one is declared; this is never something to diagnose as a false negative in the tool. If the limit's number was simply wrong for what this metric should hold going forward — not because of this change, but as a standing decision — that is a deliberate edit to make and explain on its own, never a side effect of getting one check to pass. ## Staleness `--check reports` compares a committed report against a fresh measurement and fails when they disagree, whether or not any limit is involved. Two causes produce this, and only one of them is a real regression: - **The code actually changed** and the committed report was never re-measured. Re-run with `--output-json`/`--output-markdown` to refresh it. - **The runtime changed.** Compressed sizes are Node-version dependent — the bundled zlib differs release to release — so a report written on one Node version and checked on another reads as stale with zero code changes involved. Check on the exact runtime the repository pins for everything else before concluding the source moved. A **breach** and **staleness** are reported as separate findings even when they come from the same run; do not read one as a symptom of the other. ## A run that failed outright Distinct from both of the above: the run itself could not finish, collected into `failures` in the report rather than folded into any metric. - **A limit bound to nothing.** The metric path names no measurement the run actually took — misspelled, or naming an analysis its input never ran, or genuinely ambiguous between two inputs. Fix the path in the configuration; the `codometer-configure` skill covers how a metric path resolves and why an ambiguous one is refused rather than guessed. - **An input matched no files while carrying a limit.** Declaring a limit against an input asserts its files exist, so an empty match means the glob stopped matching or the build that should have produced those files never ran — check the build step before touching the configuration. An input nobody limited that matches nothing is unremarkable and reports no failure. - **The configuration itself was rejected before anything measured.** A missing `format`, two `outputs` entries of the same type, a custom statistic naming none of `patterns`/`symbols`/`comment`, or an `outputs` entry selecting a `custom` label the top-level `custom` array never declared each fail the run at load time with the reason named. A configuration exporting a function instead of a plain object is a quieter version of the same problem: there is no config-as-function escape hatch any more, so it is not recognized and falls back to an empty configuration — which then fails on the missing `format` rather than on anything naming the real cause. Fix the configuration rather than the command line. A run with no `--config` at all is refused before anything is measured: pass `--config <path>` naming the configuration file. Every failure in one run is collected and reported together, so treat a report naming three as three things to fix, not one flaky run to retry. ## Before reaching for `--output-json`/`--output-markdown` Remember that the two `--output-*` flags and `--check` are independent, with no inferred relationship between them — the `codometer-measure` skill has the full flag table. Combining an `--output-*` flag with `--check reports` is refused outright: nothing can be stale in a report the same run just wrote, so that combination is a configuration mistake to fix, not a way to force a report current. `--inputs [globs...]` is independent of `--check` too — narrowing a `measure` run to exactly the given globs does not change whether it gates or writes. There is no `--directory` on `measure`: a run always measures the process's own working directory. The `configuration` subcommand is the one place `-d, --directory` still exists, for listing what a directory's configuration resolves to rather than for measuring it.
© organizzolini · MIT · organizzolini/codebase