Enforce a /clear workflow and stop-hook guardrail in Claude Code.
npx skills add https://github.com/tianjianjiang/smith --skill smith-ctx-claude
Please help me install this Agent Skill. Skill: smith-ctx-claude Source: https://github.com/tianjianjiang/smith/tree/main/smith-ctx-claude Command: npx skills add https://github.com/tianjianjiang/smith --skill smith-ctx-claude
Claude Code context management with a pragmatic stop-hook and persistence strategy to prevent context-loss during long sessions.
Activate this skill in Claude Code to automatically enforce a context-safe /clear workflow when your context approaches the configured thresholds.
High-intent search queries and answers about installing and using this skill.
To prevent context loss in Claude Code during long sessions, use a stop-hook guardrail to enforce a context-safe /clear workflow when usage approaches critical thresholds, persisting state to Serena memory before clearing.
Serena memory persistence works with the Claude Code clear workflow by saving session state to CLAUDE.md and Serena memory before executing /clear, enabling seamless work resumption via read_memory after the context is cleared.
Yes, you need jq installed as a dependency to manage Claude Code context thresholds with a stop-hook, as the skill scripts require jq to parse context usage data and evaluate configured critical limits.
The best way to safely clear context in Claude Code without losing uncommitted work is applying a one-shot stop-hook guardrail that automatically persists state to memory before enforcing the /clear command at critical thresholds.
You should not use an automated context clear workflow in Claude Code for single-session, short-duration tasks where context limits are never approached, as the stop-hook guardrail and memory persistence overhead provides no operational benefit.
kurokeitaCodify rules for navigating, searching, and editing code with Serena MCP tools.
maxwell2732Coordinate long-running task loops with fresh context and controlled termination.
oozoofrogAnalyze a repository and scaffold hierarchical CLAUDE.md and AGENTS.md context files.
tarrragonPrompt manual context compaction at strategic workflow milestones via a PreToolUse hook.
