Skip to content

imported-repo-skills.md

Latest commit

 

History

History
145 lines (120 loc) · 4.97 KB

imported-repo-skills.md

File metadata and controls

145 lines (120 loc) · 4.97 KB
Raw

Repository Skills Catalog

This catalog describes the repository-skill collection currently published in the AREX-Skill Library under skills/repositories/. The collection contains exactly 1,000 repository skill roots and is routed by the fixed area -> family taxonomy.

Scope

Metric Value
Repository skill roots 1,000
Taxonomy areas 20
Taxonomy families 178
Area-family memberships 2,209
Multi-assigned repositories 700
Unclassified / blocked / failed 0 / 0 / 0

A repository can appear in multiple families when it provides distinct, substantial capabilities supported by repository evidence. The router is not a replacement for a repository skill; it progressively narrows a request from an area to a family and then to the selected repository-skill root.

For a human-readable inventory of every repository grouped by area and family, see the Repository Catalog. It links each entry to the corresponding repository skill root and source repository.

Published layout

skills/
  repositories/
    repo-skills/
      repository-index.jsonl
      <skill-id>/
        SKILL.md
        sub-skills/
        references/
        scripts/
    repo-skills-router/
      SKILL.md
      references/
        areas/
        families/
        index/

The machine-readable files are the source of truth for the catalog and router:

The router's generated Markdown views are available under repo-skills-router/references/areas/ and repo-skills-router/references/families/. Use those pages for progressive disclosure; use the JSONL indexes for exact machine-readable membership.

Area summary

Area Families Memberships
Computer Vision 21 312
Biomedical AI 11 50
Generative Media 13 194
Speech and Audio 5 56
Natural Language Processing 9 78
LLM Applications 16 325
LLM Models, Training, and Alignment 6 149
Information Retrieval 4 67
MLOps 15 116
Model Deployment and Optimization 4 114
Training Infrastructure 12 135
Reinforcement Learning 5 82
Robotics and Embodied AI 7 81
Autonomous Driving 3 38
Graph Learning 4 42
Scientific Computing 16 124
Data Science 12 152
Time Series Analysis 7 53
Probabilistic and Causal Modeling 4 20
Responsible AI 4 21

The family-level membership lists are intentionally generated rather than duplicated in this document. Open the relevant area/family page or query assignments.jsonl when an exact repository list is needed.

Adding or changing a repository skill

Use create-repo-skill and verify-repo-skill to produce and verify the runtime graph. Classification is a separate post-verification step against the fixed taxonomy:

verified repository skill
  -> exact area-family classification
  -> external routing decision with evidence
  -> minimal v2 repo-routing-metadata.json
  -> transactional import
  -> regenerated router indexes and Markdown views

The runtime metadata file contains only the canonical owner/repository identity, skill ID, taxonomy hash, routing status, and exact assignments. Full rationale and evidence remain outside the runtime graph, preferably under <repo-path>/skills/disco/routing_decision/ and in the production audit artifacts.

If no exact family applies, use unclassified and ask the user whether to import the skill. Importing an unclassified skill requires a user-approved taxonomy extension before the normal import flow. An interrupted or inaccessible classification is blocked or failed, not an inferred route.

Do not hand-edit generated router Markdown. Use the verified importer or the router updater under its shared import lock.

Installation

The recommended installation is managed by DisCo:

disco repo-skills install
disco repo-skills status
disco repo-skills update

The managed destination is:

~/.disco/agent/skills/repositories/
  repo-skills/
  repo-skills-router/

Installation, updates, and routing use only the current skills/repositories/ layout.