eva-project-seed

What problem does it solve? Working on a large multi-tenant monorepo like EVA forces AI agents to re-scan the entire codebase on every task, wasting context tokens and losing track of recent changes, pending work, and architectural invariants. ## Core Features & Use Cases - Compact Project Map: Loads a terse reference of the EVA architecture, including NestJS modules, Supabase migrations and RLS rules, dashboard routes, and hard rules like org_id tenancy and approval flows. - Mandatory Change Logging: Appends a structured change entry plus a pending/improvement note to the seed change log after every task via the update_seed.py script. - Closed-Loop Documentation: Keeps docs/architecture.md, backlog.md, and the seed synchronized so documentation never drifts from the code. - Use Case: Before modifying the eva-core agent loop or adding a Supabase migration, load this seed to learn tenancy rules and recent changes, implement the change, then record the outcome and next improvement in the log. ## Quick Start Ask the agent to load the eva-project-seed skill before making any change to the EVA codebase and to update the seed log when finished.