skill-importer

name	skill-importer
description	Import external skills from Git repositories, local directories, or web URLs into the local skills library, adapting them to house conventions, auditing, and suggesting cross-skill integrations. Use when importing a skill, bringing in an external skill, or adding a skill from GitHub.
version	1.0.0
kind	hybrid
triggers	["import a skill","import skill from","bring in skill","add skill from github"]
intent	meta
created_at	2026-07-31T00:00:00.000Z
updated_at	2026-07-31T00:00:00.000Z
guardrails	["Always read foundation skills (skill-creator, skill-manager, dotfiles-management) before inspecting the candidate skill.","Present matching skills, compatibility conflicts, and uniqueness assessment to the user before modifying any skill files.","Require user confirmation before saving the adapted skill to the dotfiles repository.","Always run skill-manager audit on the imported skill after deployment and resolve any failing checks."]
resources	["<SKILL_PATH>/scripts/fetch.sh","<SKILL_PATH>/references/adaptation-rules.md","<SKILL_PATH>/references/cross-skill-matrix.md"]
tools	["bash","git","curl"]

Skill Importer

Import external skills from Git repositories, local folders, or remote web URLs into the local agent skill library, ensuring full compliance with house conventions, error-free auditing, and cross-skill lifecycle integration.

Workflow

Step 1 — Load House Foundation Skills

Before reading the candidate skill, inspect our core skill foundation to align with current ecosystem conventions:

1. View skill-creator: skill-creator/SKILL.md & skill-creator/references/house-conventions.md.
2. View skill-manager: skill-manager/SKILL.md.
3. View dotfiles-management: dotfiles-management/SKILL.md.

Step 2 — Fetch & Read Candidate Skill

1. Execute the fetch helper script to stage the candidate skill in a temporary workspace:

   <SKILL_PATH>/scripts/fetch.sh <source> [subpath]
   

2. Read the staged skill's SKILL.md, scripts, references, and manifest files thoroughly.

Step 3 — Perform Ecosystem & Skill-Creator Impact Analysis

Evaluate the candidate skill against skill-creator conventions across five dimensions:

1. Existing Overlap: Check AGENTS-TEMPLATE.md or ~/dotfiles/workspace/aistuff/skills/ for matching or conflicting skills.
2. Skill-Creator Kind Classification: Classify as guidance (prose only, zero scripts), pipeline (thin SKILL.md + scripts/ + summary projection), or hybrid.
3. Data-App / Mosaic Dashboard Assessment: Determine if the skill produces visual data/reports that should be backed by a Mosaic dashboard tile (references/data-app-skills.md). Check if webapp/app.json exists or if adding a webapp tile is beneficial.
4. House Conventions & Conflicts: Check YAML frontmatter, line budgets (ceiling 200 lines—if >200 lines, extract into references/<topic>.md preserving 100% of original content intact), environment paths ($WORKSPACE_PATH, $SCRIPTS_PATH, $TOOLS_PATH, <SKILL_PATH>), executable tools list, and script output contracts.