What problem does it solve? When maintaining a long-lived codebase like FutureFin, developers repeatedly risk re-proposing designs that were already investigated and rejected, or reintroducing bugs that were already fixed. This Skill provides the historical chronicle of every major investigation, dead end, and removal as symptom → root cause → evidence → status, so no future session re-fights a settled battle. ## Core Features & Use Cases - Settled battles table: A catalog of 23 rejected or removed approaches (age-based retirement triggers, per-asset contribution config, external-database mode, OAuth login, and more) with commit hashes, versions, and migration files as evidence. - Detailed forensic entries: In-depth write-ups of root causes, such as the table-CSS saga, inverted SQL binds between Household/Mine branches, hybrid-density deflation bugs, and the five traps of embedding PostgreSQL in the Docker image. - Use Case: Before proposing to add a density parameter to the get_projection MCP tool or to make the container honour DATABASE_URL again, load this Skill to learn these were deliberate non-additions or scheduled removals, with the reasoning and CI guards that pin them. ## Quick Start Load the futurefin-failure-archaeology skill before proposing any change to FutureFin's projection engine, Docker image, authentication, or database topology to check whether the idea was already rejected.