Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
AI_SKILLS_HOMENoDefines an alternative directory to use instead of the default ~/.ai-skills local environment.

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

Roteador de skills locais. Antes de tarefas de desenvolvimento, arquitetura ou automação, chame search_installed_skills e carregue com get_skill_content apenas as skills escolhidas. Novas skills só entram como rascunho (propose_skill_draft) e só são promovidas (promote_draft) com aprovação explícita do humano.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}
logging
{}
prompts
{
  "listChanged": true
}
resources
{
  "subscribe": false,
  "listChanged": true
}

Tools

Functions exposed to the LLM to take actions

NameDescription
search_installed_skillsA

Busca skills instaladas por relevância e devolve só metadados resumidos + score.

Não carrega nenhum SKILL.md. Use palavras-chave da stack/tarefa (os metadados costumam estar em inglês) e faça mais de uma busca se a tarefa cruzar domínios.

get_skill_contentA

Lê o SKILL.md completo de UMA skill escolhida (use após search_installed_skills).

rebuild_indexB

Força a reindexação completa do diretório oficial de skills.

propose_skill_draftA

Salva um RASCUNHO de skill em drafts//SKILL.md (nunca em skills/).

Só chame depois de o humano confirmar explicitamente que quer o rascunho. content é o corpo em Markdown; um frontmatter enviado junto é mesclado.

promote_draftA

Move drafts/<skill_name> para skills/ e reindexa. Exige aprovação humana explícita.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4/5.0

Scored across 5 tools

Disambiguation5/5

Each tool targets a clearly distinct action: search (discover), get content (read), rebuild (maintenance), propose draft (create), promote draft (publish). The descriptions even specify sequencing (use search before get, propose before promote), leaving no room for misselection.

Naming Consistency5/5

All five tools follow a consistent snake_case verb_noun pattern: search_installed_skills, get_skill_content, rebuild_index, propose_skill_draft, promote_draft. No mixed conventions.

Tool Count5/5

Five tools map cleanly onto the discover/read/maintain/draft/publish lifecycle of a skills router. Well within the ideal 3-15 range with no filler.

Completeness4/5

The core discovery and draft-promotion lifecycle is covered, but there is no discard/delete draft, edit-existing-skill, or list-drafts operation, so agents may hit dead ends when managing or reverting drafts.

Maintenance

ActivityMaintained
ResponsivenessNo recent community issues