Decisions, fixes, patterns — indexed and searchable by your AI agent mid-session. Wardwell is an MCP server that turns your accumulated knowledge into tools the agent calls on its own.
Decisions, history, blockers, lessons — written to a plain-text vault as you go. Full-text indexed automatically. Markdown and JSONL you can read yourself.
Wardwell exposes search as MCP tools. When your agent needs context — a past decision, a pattern from another project, that fix from three weeks ago — it queries the index directly.
Nothing stays siloed. A lesson learned in one project surfaces in another. Your vault is one search space — every project, every session, every insight.
Links, backlinks, and entity resolution across your vault. The agent traverses relationships in ~200 tokens instead of reading full files. Wiki-links, callsigns, and frontmatter refs — all indexed automatically.
Partial reads with offset and limit. The agent checks file size before committing tokens — 838-line PRD? Read lines 100–105. Frontmatter always comes back in full.
Feature flags in config.yml. Every new capability defaults on, flips off with one line. Graph navigation, entity resolution, partial reads — toggle each independently.