Skip to content

Releases: SantanderAI/ralph-vault-skill

v0.1.0 — Initial public release

16 Jun 18:22

Choose a tag to compare

First public release of ralph-vault by Santander AI Lab: a Python CLI tool / library that creates and maintains a progressive-disclosure knowledge vault (a tiered "deepwiki") for one or many code repositories — the knowledge source for LLM / AI agent loops (ralph-style).

Added

  • ralph-vault agent skill: project-agnostic, progressive-disclosure knowledge vault for one or many code repositories.
  • scripts/gv.py — stdlib-only CLI: init, add, delete, list, check, validate, plan, mark-synced, changelog.
  • Tiered vault structure (index, domains, repos, components, infrastructure, technologies, relations, cross-cutting, adrs, glossary, agent-context, meta).
  • Per-repo registry in .ralphvault/config.json (no hardcoded repos), path-aware staleness detection and module-level incremental sync.
  • Deterministic validation gate (frontmatter, wikilinks, token budgets, no-source-code rule, source-backlink checks) and FIXED prompts for ralph-loop content generation.
  • install.sh for Codex/Claude/Gemini skill directories.

Licensed under Apache-2.0. See CHANGELOG.md for details.