Ejecuta cualquier Skill en Manus
con un clic

Ejecuta cualquier Skill en Manus con un clic

verify-skill

Estrellas4

Forks1

Actualizado16 de julio de 2026 a las 16:02

Use when validating a new or changed Claude Code skill, checking structural and repository policy compliance, testing whether descriptions trigger accurately, or grading representative skill outputs before deployment.

Instalación

Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.

Ejecutar en Manus

Fuente

alvis

alvis/.claude

Abrir repositorio de GitHub Ver repositorios del creador

Descarga

Ejecutar en Manus

Ocupaciones relacionadasSOC

Basado en la clasificación ocupacional SOC

Analistas de garantía de calidad de software y probadoresOcupaciones informáticas y matemáticas·SOC 15-1253

Explorador de archivos

12 archivos

SKILL.md

readonly

name	verify-skill
description	Use when validating a new or changed Claude Code skill, checking structural and repository policy compliance, testing whether descriptions trigger accurately, or grading representative skill outputs before deployment.
model	opus
context	fork

Verify Skill

Boundaries

Claude Code's validator owns manifest and frontmatter schema correctness.
quick_validate.py owns repository policies only; it must not duplicate the evolving Claude schema.
Functional and trigger checks are required when behavior or discovery changed, not for a wording-only edit with unchanged meaning.

Shared authoring policy lives at ${CLAUDE_SKILL_DIR}/../../constitution/references/authoring-invariants.md.

Inputs

Required: a SKILL.md, skill directory, plugin, marketplace, or plugins path.
Optional mode: structural, functional, or full (default full).
Optional evaluation data: evals/evals.yaml and representative prompts.

Workflow

Resolve the target and enumerate affected skills and plugin roots.
Resolve each affected plugin root from the target path, then run official strict validation for every affected plugin. Report Claude's errors verbatim enough to identify their source; do not reinterpret valid fields through a private schema.
Run ${CLAUDE_SKILL_DIR}/scripts/quick_validate.py on the target. Review body length, description budget, unresolved local references, and placeholders.
For functional or full mode, load existing evals or derive a small set from the owned outcome. Follow references/functional-mode.md for execution and grading details.
Test positive trigger prompts and nearby negative prompts. A pass requires the intended skill to trigger without stealing work from its neighbors.
Aggregate evidence by skill. When fixes are requested, change only reported causes and rerun the failed checks.

The validator does not require fixed headings, personas, diagrams, subagent prompts, XML report envelopes, or a section literally named “Skill Completion.”

Verification

claude plugin validate --strict <plugin-path>
python3 "${CLAUDE_SKILL_DIR}/scripts/quick_validate.py" <skill-or-plugin-path>
python3 "${CLAUDE_SKILL_DIR}/scripts/run_trigger_eval.py" --help

Run the first two commands after every fix iteration. Trigger and functional results must include the prompt, observed behavior, and pass/fail rationale.

Completion

Return a concise per-skill verdict with official validation status, policy issues, functional/trigger evidence when applicable, and actionable failures.

Más de este repositorio

mismo repositorio

create-agent

alvis/.claude

Creates a new specialist agent as two stitched source files, base.md plus frontmatter/claude.json, proposing model, effort, and permissions by role archetype and confirming them with the user before writing. Use when adding a new subagent, defining a new specialist role, scaffolding an agent definition, or when update-agent hands off new-agent creation.

2026-07-184

install-agents

alvis/.claude

Discover, validate, stitch, and install specialist agent templates contributed by Essential and other enabled plugins in the same marketplace. Use when asked to install agents, set up subagents, refresh the agent team, or configure Claude Code on a new machine.

2026-07-174

skill-name

alvis/.claude

[Description] — third person, what the skill does plus a 'Use when...' clause with concrete trigger phrases a user would actually type; 25-60 words; lead with the key use case.

2026-07-174

update-agent

alvis/.claude

Update explicitly selected agent definitions to the current two-file template or a stated behavior change while preserving useful role expertise, trigger boundaries, context, collaboration links, and working voice. Use when migrating agents to a template revision, correcting agent configuration, or batch-updating selected agents; require an exact selector and route genuinely new roles to create-agent.

2026-07-164

lint

alvis/.claude

Enforce coding standards mechanically across a selected scope with batched linters and independent reviewers. Use when source files need lint-error correction, standards enforcement, or consistent formatting, including calls extended by another plugin's portable lint profile; behavior-changing repairs belong to fix.

2026-07-164

review-code

alvis/.claude

Review semantic correctness, security, test intent, documentation, sibling consistency, and alignment with the implementation plan. Use after code changes or for explicit review requests; report findings without editing code and leave mechanical standards enforcement to lint.

2026-07-164

Verify Skill

Boundaries

Claude Code's validator owns manifest and frontmatter schema correctness.

quick_validate.py owns repository policies only; it must not duplicate the evolving Claude schema.

Functional and trigger checks are required when behavior or discovery changed, not for a wording-only edit with unchanged meaning.

Shared authoring policy lives at ${CLAUDE_SKILL_DIR}/../../constitution/references/authoring-invariants.md.

Inputs

Required: a SKILL.md, skill directory, plugin, marketplace, or plugins path.

Optional mode: structural, functional, or full (default full).

Optional evaluation data: evals/evals.yaml and representative prompts.

Workflow

Resolve the target and enumerate affected skills and plugin roots.

Resolve each affected plugin root from the target path, then run official strict validation for every affected plugin. Report Claude's errors verbatim enough to identify their source; do not reinterpret valid fields through a private schema.

Run ${CLAUDE_SKILL_DIR}/scripts/quick_validate.py on the target. Review body length, description budget, unresolved local references, and placeholders.

For functional or full mode, load existing evals or derive a small set from the owned outcome. Follow references/functional-mode.md for execution and grading details.

Test positive trigger prompts and nearby negative prompts. A pass requires the intended skill to trigger without stealing work from its neighbors.

Aggregate evidence by skill. When fixes are requested, change only reported causes and rerun the failed checks.

The validator does not require fixed headings, personas, diagrams, subagent prompts, XML report envelopes, or a section literally named “Skill Completion.”

Verification

claude plugin validate --strict <plugin-path> python3 "${CLAUDE_SKILL_DIR}/scripts/quick_validate.py" <skill-or-plugin-path> python3 "${CLAUDE_SKILL_DIR}/scripts/run_trigger_eval.py" --help

Run the first two commands after every fix iteration. Trigger and functional results must include the prompt, observed behavior, and pass/fail rationale.

Completion

Return a concise per-skill verdict with official validation status, policy issues, functional/trigger evidence when applicable, and actionable failures.