Hook Claude Code
Runs when you submit a prompt, before the agent sees it, executing user-prompt-submit.sh via bash. From roboco-io/ralph-mem.
Hook Claude Code
Runs when you submit a prompt, before the agent sees it, executing user-prompt-submit.sh via bash. From roboco-io/ralph-mem.
Settings file Claude Code
Agent settings declaring 1 hook event (UserPromptSubmit).
Instructions file
Project instructions for ralph-mem, a Claude Code plugin that stores and retrieves work context across sessions. It combines repeated task attempts with memory management to reduce lost context and declining responses as conversations grow.
Command
A command for deleting saved memory entries from selected sessions, dates, or categories.
Command
A command for manually saving information in the agent's persistent memory, optionally with a category and tags.
Command
A command for searching saved memory records and session information by keyword. It can show either a short result list, timeline context, or full details for a selected record.
Command
A command for viewing the current state and statistics of the saved-memory system.
Command
A command set for repeatedly running a task until a chosen condition is met, such as passing tests, building successfully, or having no lint errors.
Hook
Runs after a tool call finishes, executing post-tool-use.js via bun. From roboco-io/ralph-mem.
Hook
Runs when the agent finishes a response, executing stop.js via bun. From roboco-io/ralph-mem.
Skill Claude CodeCodex
A command for removing selected records from stored memory.
Skill Claude CodeCodex
A skill for manually saving project context in persistent memory, with optional categories and tags.
Skill Claude CodeCodex
A command for searching saved memory records and session information, with results revealed from short summaries to full details.
Skill Claude CodeCodex
A command that reports the current state of the memory system, including saved sessions, observations, storage use, and recent activity.
Skill Claude CodeCodex
A command for repeating a task until a chosen success check passes, such as tests, a build, linting, or type checking.