Persist and organize durable project memories in a centralized store.
npx skills add https://github.com/BumpyClock/dotfiles --skill memories
Please help me install this Agent Skill. Skill: memories Source: https://github.com/BumpyClock/dotfiles/tree/main/.ai_agents/skills/memories Command: npx skills add https://github.com/BumpyClock/dotfiles --skill memories
Capture, update, and curate durable project memories such as architecture decisions, coding preferences, workflows, tooling choices, and domain rules. Use when the user asks to save/remember a memory or when you identify high-leverage knowledge that should persist across sessions. Also use when asked to review, validate, or prune stored memories.
Tell the AI to save a memory to the memories repository.
High-intent search queries and answers about installing and using this skill.
To save architecture decisions across sessions, tell the AI to save a memory to the centralized memory store. The skill automatically captures, tags, and indexes durable project knowledge for quick recall and reuse.
Organizing project memories like coding preferences and tooling choices is handled by applying automated tagging, categorization, and indexing in a centralized store. The skill manages metadata using YAML frontmatter to maintain a structured knowledge base.
You can review and prune outdated domain rules by asking the AI to validate or clean up stored memories. This process ensures the persisted knowledge base remains relevant and durable over time.
Memory evolution for tracking workflow and tooling changes works by storing metadata in YAML frontmatter and maintaining an evolution history within the centralized memory store. This allows updates to existing memories while preserving their historical context.
No specific dependencies are required to manage architecture decisions. The skill operates independently using internal scripts and references to handle the creation, updating, and indexing of memories in the centralized store.
You should avoid a centralized memory store for transient or non-reusable project knowledge. The approach is designed specifically to preserve high-leverage, durable rules and decisions across multiple sessions and project reviews.
pradeepmouliGenerate baseline-spec.md and current-state.md from repository analysis.
siddsacharAudit memory vaults for duplicates, stale entries, and broken links.
winsorllcStore, search, and retrieve memories across jobs using JSON files.
wshobsonCreate and manage Architecture Decision Records using standardized templates.
