Claude Skill

skills-help

The skills library menu — every skill grouped by family, each with a one-line description and a dual-doc/single-doc marker; pass a family name to drill into one family in full detail. Trigger with /skills-help, "list my skills", "which skill for X", "skills menu", "what skills do

LLM Mart · 0 points · 0 views 0 listing impressions 0 install-command copies
Download allemaar-open-skills-skills_skills-help-acf5f79.zip · 3 KB

Install

skills CLI npx skills add https://github.com/allemaar/open-skills/tree/main/skills/skills-help
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install allemaar-open-skills@llmmart
Git git clone https://github.com/allemaar/open-skills.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole allemaar/open-skills collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

/skills-help

The live menu for the skills library. Family membership comes from the bundled taxonomy.yon; names, descriptions, triggers, runtime declarations, and protocol presence come from the installed sibling skill folders. This is a reference card and changes nothing.

How to use

  • /skills-help — print every installed skill grouped by the taxonomy's family order.
  • /skills-help {family} — resolve the argument against family ID, label, or aliases and print only that family in full detail.
  • /skills-help {need} — when the argument is not a family, match the need against live descriptions and trigger phrases and show the closest routes without inventing capabilities.

Build the live view

  1. Locate the active skills root containing this skills-help folder. Enumerate only its immediate sibling directories that contain SKILL.md.
  2. Read taxonomy.yon. Its family:* records define IDs, labels, aliases, and display order; SkillFamilies assigns the skills shipped by open-skills.
  3. Intersect taxonomy assignments with the installed siblings. Do not show an absent pack skill as installed.
  4. For every installed sibling, read current SKILL.md frontmatter for name, description, triggers, and optional runtime. Detect protocol.yon directly from that sibling folder.
  5. Put an installed sibling absent from SkillFamilies under Unclassified with its source-derived description. Do not guess a best-fit family and do not modify the taxonomy.
  6. If taxonomy parsing fails, report that the categorized menu is unavailable and list installed skill names alphabetically under Unclassified. Do not silently fall back to the old embedded roster.

Render

  • In full-menu mode, show each family label followed by one line per installed skill: marker, name, and the first complete sentence of its live description.
  • In family mode, show name, full live description, trigger phrases, format, and runtime. Runtime defaults to [claude, codex, agents] when absent.
  • In need-matching mode, name the matched phrases that support each route. If nothing matches, say so and offer /skill-smith or /insight-skill-gap rather than fabricating a recommendation.
  • Keep Unclassified last. State that it means “installed but not assigned by this pack's taxonomy,” not “unsafe” or “low quality.”

Marker legend:

  • ◆ SKILL.md plus a declarative, mechanically validatable protocol.yon
  • ◇ SKILL.md without a protocol companion

Optional journeys

When someone asks where to start, offer a proportionate next step using the live view above. A journey is advice, not an execution chain. Do not add a lifecycle to a simple menu request, change menu formats or require workers.

Resolve suggestions against installed siblings and current descriptions. Name the actual trigger that fits. Upstream, draft and absent skills are not installed; describe the useful activity in prose when no available owner matches. Do not invent registered invocations.

Examples, adapted to actual availability:

  • Fuzzy idea: clarify the decision and blocking unknown; suggest an interview only if available and wanted.
  • Bug: gather reproduction and evidence, then choose focused diagnosis. A supplied fix may instead need review.
  • Change or review: distinguish planning, approved-plan execution and inspecting an actual candidate. Offer only the needed stage.
  • Writing: distinguish drafting, repairing one text and combining several reports; suggest the matching installed capability.

Offer a further step only when useful, with material costs or prerequisites. The caller chooses whether to continue. This skill still reads and reports only.

Boundary

This skill reads installed metadata and reports it. It does not install, edit, classify, or execute other skills. To scaffold a new skill use /skill-smith.

Files (open-skills)
  • SKILL.md 4.7 KB
    ---
    name: skills-help
    description: The skills library menu — every skill grouped by family, each with a one-line description and a dual-doc/single-doc marker; pass a family name to drill into one family in full detail. Trigger with /skills-help, "list my skills", "which skill for X", "skills menu", "what skills do I have". Not /skill-smith (which scaffolds a new skill).
    visibility: public
    triggers:
      - "/skills-help"
      - "/skills-help {family}"
      - "list my skills"
      - "which skill for X"
      - "skills menu"
      - "what skills do I have"
    next-skills:
      - skill: skill-smith
        phrase: "/skill-smith"
        why: "When the library menu shows no skill covers the need, create the missing one."
      - skill: insight-skill-gap
        phrase: "/insight-skill-gap"
        why: "After browsing the roster and finding a hole, run the gap analysis to decide what to add or update."
    ---
    
    # /skills-help
    
    The live menu for the **skills library**. Family membership comes from the bundled [`taxonomy.yon`](taxonomy.yon); names, descriptions, triggers, runtime declarations, and protocol presence come from the installed sibling skill folders. This is a reference card and changes nothing.
    
    ## How to use
    
    - **`/skills-help`** — print every installed skill grouped by the taxonomy's family order.
    - **`/skills-help {family}`** — resolve the argument against family ID, label, or aliases and print only that family in full detail.
    - **`/skills-help {need}`** — when the argument is not a family, match the need against live descriptions and trigger phrases and show the closest routes without inventing capabilities.
    
    ## Build the live view
    
    1. Locate the active skills root containing this `skills-help` folder. Enumerate only its immediate sibling directories that contain `SKILL.md`.
    2. Read `taxonomy.yon`. Its `family:*` records define IDs, labels, aliases, and display order; `SkillFamilies` assigns the skills shipped by open-skills.
    3. Intersect taxonomy assignments with the installed siblings. Do not show an absent pack skill as installed.
    4. For every installed sibling, read current `SKILL.md` frontmatter for name, description, triggers, and optional runtime. Detect `protocol.yon` directly from that sibling folder.
    5. Put an installed sibling absent from `SkillFamilies` under **Unclassified** with its source-derived description. Do not guess a best-fit family and do not modify the taxonomy.
    6. If taxonomy parsing fails, report that the categorized menu is unavailable and list installed skill names alphabetically under **Unclassified**. Do not silently fall back to the old embedded roster.
    
    ## Render
    
    - In full-menu mode, show each family label followed by one line per installed skill: marker, name, and the first complete sentence of its live description.
    - In family mode, show name, full live description, trigger phrases, format, and runtime. Runtime defaults to `[claude, codex, agents]` when absent.
    - In need-matching mode, name the matched phrases that support each route. If nothing matches, say so and offer `/skill-smith` or `/insight-skill-gap` rather than fabricating a recommendation.
    - Keep **Unclassified** last. State that it means “installed but not assigned by this pack's taxonomy,” not “unsafe” or “low quality.”
    
    Marker legend:
    
    - **◆** `SKILL.md` plus a declarative, mechanically validatable `protocol.yon`
    - **◇** `SKILL.md` without a protocol companion
    
    ## Optional journeys
    
    When someone asks where to start, offer a proportionate next step using the live view above. A journey is advice, not an execution chain. Do not add a lifecycle to a simple menu request, change menu formats or require workers.
    
    Resolve suggestions against installed siblings and current descriptions. Name the actual trigger that fits. Upstream, draft and absent skills are not installed; describe the useful activity in prose when no available owner matches. Do not invent registered invocations.
    
    Examples, adapted to actual availability:
    
    - Fuzzy idea: clarify the decision and blocking unknown; suggest an interview only if available and wanted.
    - Bug: gather reproduction and evidence, then choose focused diagnosis. A supplied fix may instead need review.
    - Change or review: distinguish planning, approved-plan execution and inspecting an actual candidate. Offer only the needed stage.
    - Writing: distinguish drafting, repairing one text and combining several reports; suggest the matching installed capability.
    
    Offer a further step only when useful, with material costs or prerequisites. The caller chooses whether to continue. This skill still reads and reports only.
    
    ## Boundary
    
    This skill reads installed metadata and reports it. It does not install, edit, classify, or execute other skills. To scaffold a new skill use `/skill-smith`.
    
  • taxonomy.yon 4.8 KB · in bundle

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.