solana-foundation/solana-com · Archived

llms-txt-generator

Upgrade llmtxt-generator.py by scanning apps/ for doc structure changes, then regenerate llms.txt and llms-en.txt while only adding missing sections. Use when updating the generator, adding new doc sections, or refreshing LLM text files.

First seen Apr 14, 2026

Installation

$ npx skills add solana-foundation/solana-com --skill llms-txt-generator

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from solana-foundation/solana-com.

npx skills add solana-foundation/solana-com

Browse all from solana-foundation/solana-com

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 486
License LICENSE
Default branch main
Open issues 30
Status Archived

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,545 B
  • docs SUMMARY.md 263 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 5 installs

SKILL.md

LLMs.txt Generator Upgrader

Goal

Upgrade llmtxt-generator.py using current app structure and existing apps/web/public/llms.txt, then regenerate outputs. Only add new or missing sections and links. Do not remove, rename, or rewrite existing sections unless the user explicitly asks.

Discovery: apps structure

  1. Scan apps/ to understand doc surfaces and routing structure.
  2. Note any new top-level doc categories or developer resources that are not

represented in the current llms.txt.

Baseline: current llms.txt

  1. Read apps/web/public/llms.txt.
  2. Extract section headings (##) and link lists.
  3. Treat this as the canonical baseline: preserve all existing sections and

links, and avoid duplicates.

Update the generator (llmtxt-generator.py)

Always edit the generator, not the output files.

  • Add only missing sections or links to CURATEDSECTIONS or ENGLISHEXTRAS

so the generator reproduces the current llms.txt plus any new additions.

  • Keep descriptions short, factual, and consistent with the current style.
  • Keep URL patterns canonical (no utm_ parameters).
  • Keep LOCALES aligned with packages/i18n/src/config.ts and update

LANGUAGE_NAMES for any newly added locale.

  • If new routing patterns appear, update normalizellmsurls() accordingly.

Regeneration

From the repo root:

python llmtxt-generator.py

Outputs:

  • apps/web/public/llms.txt
  • apps/web/public/llms-en.txt

Output constraints

  • Only add new or missing sections/links; never delete existing ones.
  • Preserve ordering unless a new section clearly belongs next to related ones.
  • Do not fabricate llms-full.txt. Follow repo guidance or user instructions.
  • For non-English locales, note the generator preserves existing content if the

file exists; change this behavior only if explicitly requested.

Verification

  • apps/web/public/llms.txt contains all prior sections and any new additions.
  • Locale files exist for all configured locales.
  • Output URLs are canonical and contain no utm_ parameters.

Example requests

  • "Upgrade llmtxt-generator.py after a new docs section"
  • "Add missing llms.txt sections based on apps structure"
  • "Add a new locale and regenerate llms files"