From reload-and-reset
Updates the AI instruction system to the latest version and resets AI behavior to follow instructions
How this skill is triggered — by the user, by Claude, or both
Slash command
/reload-and-reset:reload-and-resetThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
> **Language:** Respond in the user's language. If unclear, default to the language of the user's message.
Language: Respond in the user's language. If unclear, default to the language of the user's message.
Updates the AI instruction system to the latest version and corrects AI behavior to properly follow instructions.
Save current task state
echo "Saving current task state..."
scripts/checkpoint.sh pending > /tmp/ai_tasks_backup.txt 2>&1 || echo "No task information"
Update submodule (skip if this is the project itself)
if [ -d "instructions/ai_instruction_kits/.git" ]; then
echo "Updating AI instruction system..."
git submodule update --remote instructions/ai_instruction_kits
else
echo "Running in AI instruction kit dev environment (skipping submodule update)"
fi
Verify update status
echo "Current instruction system status:"
if [ -d "instructions/ai_instruction_kits/.git" ]; then
git submodule status instructions/ai_instruction_kits
else
echo "Dev environment: $(git rev-parse --short HEAD)"
fi
System reset declaration Reset the AI system with the following procedure:
I have now been reset to the following state:
Reload ROOT_INSTRUCTION
Auto-detect path and load:
instructions/ja/system/ROOT_INSTRUCTION.md or
instructions/ai_instruction_kits/instructions/ja/system/ROOT_INSTRUCTION.md
Check saved task state
if [ -f "/tmp/ai_tasks_backup.txt" ]; then
echo "Saved tasks:"
cat /tmp/ai_tasks_backup.txt
rm /tmp/ai_tasks_backup.txt
fi
/reload-and-reset
No arguments required. The AI system will be fully reset and will operate according to the latest instructions.
npx claudepluginhub dobachi/claude-skills-marketplace --plugin reload-and-resetPerforms structured self-healing assessment across AI subsystems to identify drift, staleness, and misalignment, then rebalances through grounding and memory integration. Use mid-session when responses feel formulaic, after error chains, or as proactive maintenance.
Maintains CLAUDE.md and AGENTS.md instruction files by enforcing size limits (<300 lines), progressive disclosure via docs/ references, multi-agent compatibility, and tool-first content. Use for creation, updates, audits.
Optimizes AI-facing instructions for subagents, CLAUDE.md, rules, skills, agents by removing discoverable data, explained knowledge, invented constraints, stale facts. Use for authoring or auditing bloat.