@adelvillar1

project-knowledge-graph indexes markdown artifacts (recaps, plans, architecture docs, skills) across multiple projects into a local FalkorDB instance, enabling concept-based cross-project search with cypher filtering and tf-idf ranking.

init-project-structure scaffolds a new project with slim CLAUDE.md, contract docs, and a docs/ tree organized for plans and recaps. asks topology preference and detects existing code to avoid greenfield assumptions.

tech-stack-evaluation guides users through diagnosing maintainability, performance, and feature blockers before recommending architectural changes. it emphasizes evidence-based options over blanket rewrites and ties recommendations to codebase assessment.

slim-project-memory restructures bloated project context files into a router-style CLAUDE.md, gitignored local env reference, and topical docs tree, with a housekeeping protocol to prevent regrowth.

codebase-survey maps project structure, scale, architecture, and data model through a 14-step workflow prioritizing claude.md context and targeted domain deep dives over generic breadth-first scans.

Complete project lifecycle methodology for Hermes Agent sessions. A single, integrated workflow covering warmup → plan → build → recap → wrapup — with contex...