Skip to main content

skill-librarian

Use when a request is complex, unfamiliar, multi-part, explicitly asks for a skill or playbook, names a specialized deliverable or non-routine domain guidance need, requires unfamiliar domain guidance, or spans two or more independent domains. Skip routine work fully covered by an installed process, project, or domain skill.

Ir para a instalação

Informações da origem

Repositório
hbui290/agentic-skill-library
Última atividade na origem
22 de julho de 2026 às 12:39
Idioma detectado do SKILL.md
inglês
Estrelas
1
Forks
0

Opções de instalação

Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.

Revise os arquivos de origem

Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.

Explorador de arquivos
7 arquivos

Exibindo SKILL.md

SKILL.md
Instruções da origem · Visualização somente leitura
name
skill-librarian
description
Use when a request is complex, unfamiliar, multi-part, explicitly asks for a skill or playbook, names a specialized deliverable or non-routine domain guidance need, requires unfamiliar domain guidance, or spans two or more independent domains. Skip routine work fully covered by an installed process, project, or domain skill.
# Skill Librarian Apply an applicable Official Superpowers skill for process first; then invoke Librarian in the same task phase when triggered. ## Required trigger check Before planning or execution, invoke Librarian when: - specialized/unfamiliar guidance could prevent a wrong approach; - more than one workstream, deliverable, or domain may need composition; or - User explicitly asks for the Librarian or to find/select skills. Do not invoke it for simple general reasoning, direct edits without specialized guidance, or work covered by a native skill. Do not invoke it merely because a request mentions a tool or service; do not skip it because the task title sounds clear. ## Registry ```bash REGISTRY_ROOT="${AGENTIC_SKILL_REGISTRY_ROOT:-$HOME/.agents/agentic-skill-library}" ``` The registry CLI is the only discovery and loading runtime. Do not inspect or load catalog entries directly. ## Current-phase routing Read only the minimum reference set for the current phase. | Phase need | Reference to read | | --- | --- | | Every invoked phase: search, retry, select, and read | [control plane](references/control-plane.md) | | Every invoked phase: user-visible receipt | [decision trace](references/decision-trace.md) | | Safety metadata, action scope, or owner confirmation | [trust and safety](references/trust-and-safety.md) | | Multiple skills/phases or a Librarian subagent | [composition](references/composition.md) | | Admit a reviewed external source | [source intake](references/source-intake.md) | | Pressure-test routing or release this contract | [evaluation](references/evaluation.md) | ## Hard Rules - Never load the entire catalog or dump the whole discovery index. - Do not execute bundled scripts automatically. - Do not grant credentials or broad permissions to a selected skill. - Never bypass quarantine, path, symlink, or hash failures. - Keep official Superpowers unmodified and use it only for process guidance. - Do not add a runtime hook, MCP integration, or automatic router; this remains an on-demand skill contract.
Ver no GitHub