convez.ai
Skill
MukundaKatta

skillint-py

MukundaKatta/skillint-py

Ver no GitHub
InicialPluginScore · de 0 a 100

Lint Claude Code SKILL.md files. Python port of @mukundakatta/skillint.

Instalação

$npx convez-ai MukundaKatta/skillint-py

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
0
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 5 meses
Versão
—
Issues abertas
0
Contribuidores
1

README

A tradução para o português deste README ainda está na fila.

skillint-py

Python License: MIT

A linter for Claude Code skill files (SKILL.md). Catches malformed YAML frontmatter, missing required fields, thin descriptions, hardcoded secrets, and missing body headings before they ship.

Stdlib-only Python port of @mukundakatta/skillint. No PyYAML required -- the linter only needs to recognize the simple key: value mapping that frontmatter actually uses.

Note: there's a separate claude-skill-check package on PyPI by a different author. This package is the user's own port and uses the -py suffix to disambiguate.

Install

pip install skillint-py

Usage

from skillint import lint

result = lint("path/to/SKILL.md")

result.fatal           # True if frontmatter could not be parsed at all
for issue in result.issues:
    print(issue.line, issue.severity, issue.rule_id, issue.message)

You can also lint an in-memory source string:

from skillint import lint_source

source = """---
name: git-commit
description: Use when the user asks Claude to commit changes; handles co-author trailers.
---

## Triggers
"""
result = lint_source(source, file="<inline>")

Result shape

@dataclass
class Issue:
    rule_id: str
    severity: str           # "error" | "warning" | "info"
    message: str
    suggestion: str = ""
    line: int = 1

@dataclass
class LintResult:
    file: str
    fatal: bool
    issues: list[Issue]

Rules

Rule idSeverityChecks
missing-frontmattererrorFile doesn't start with --- YAML fence.
yaml-parse-errorerrorFrontmatter fails to parse as a mapping.
missing-requirederrorname or description is missing.
empty-requirederrorname or description is empty / whitespace.
kebab-case-nameerrorname is not kebab-case (^[a-z][a-z0-9-]*[a-z0-9]$).
overlong-namewarningname exceeds 64 characters.
short-descriptionwarning / errorDescription below 80 chars (or below 20 -> error).
no-trigger-in-descriptionwarningDescription lacks "when / use / trigger / handle / invoke" language.
no-body-headingswarningBody has no ## headings.
hardcoded-secreterrorProvider-shaped key (OpenAI, Anthropic, GitHub, AWS, Slack, Stripe, Google AI).
unknown-fieldinfoFrontmatter field outside the known set.

API differences from the JS sibling

  • Stdlib-only YAML parser handles the flat key: value mapping that frontmatter actually uses. Block scalars and tagged values are not supported -- if your frontmatter needs them, the linter falls back to a yaml-parse-error so you'll see it loud and clear.
  • No --config plugin loading -- this Python port runs the built-in rules.
  • Spec entrypoint is lint(skill_md_path) -> LintResult. lint_source covers the in-memory case.

See the JS sibling's README for the broader rationale.

Ver o original no GitHub

Bom para

Relacionados