Save and load case-specific project memory under ~/ai-memory.
npx skills add https://github.com/ymgn/dotfiles --skill context-memory-ymgn
Please help me install this Agent Skill. Skill: context-memory Source: https://github.com/ymgn/dotfiles/tree/main/codex/skills/context-memory Command: npx skills add https://github.com/ymgn/dotfiles --skill context-memory-ymgn
This Skill enables teams and agents to create a shared, case-scoped memory that persists across sessions, allowing convenient saving of specs, decisions, next steps, and logs under ~/ai-memory for collaborative project work.
Use the context-memory scripts to save and load memory. Example: python3 ~/.codex/skills/context-memory/scripts/context_save.py CASE-XYZ "Initialized memory for Case-XYZ" python3 ~/.codex/skills/context-memory/scripts/context_load.py CASE-XYZ
High-intent search queries and answers about installing and using this skill.
To share project context across multiple AI agents, you can use a shared memory system that saves and loads case-specific data under a designated local directory. This allows different agents to access the same specs, decisions, and logs during collaborative workflows.
A CASE-ID based memory system organizes project context into specific directories containing files for specs, decisions, next steps, and logs. It enables multiple agents to append timestamped entries and resume work on a specific case across different sessions.
You can save and load agent context by using slash-style commands to trigger Python scripts that write and retrieve case data. This process creates per-case directories and appends timestamped entries to maintain a traceable history of agent decisions.
Yes, you can use Python scripts to maintain AI memory logs across sessions. The scripts create per-case directories and append timestamped entries to log files, ensuring that context persists and can be loaded when resuming work.
The best way to organize cross-agent project memory is to structure it by case, separating specs, decisions, next steps, and logs into individual files within a shared directory. This append-only approach maintains history and enables quick resumption.
ecielamScaffold Claude Code plugins with standardized directory structures and manifest configuration.
timequityLog architectural decisions as ADR entries to docs/DECISIONS.md.
ricableCoordinate distributed AI agents with shared vector memory and task routing.
tiankong0101-bytePersist agent context across sessions using session state, vector search, git notes, and markdown archives.
