Menu
Use when processing pending session transcripts into daily summaries — first session of the day, on manual request, or on schedule
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
| name | synthesize |
| description | Use when processing pending session transcripts into daily summaries — first session of the day, on manual request, or on schedule |
| user-invokable | true |
Launch a subagent to process memory transcripts into daily summaries and selectively route key learnings to long-term memory.
IMPORTANT: This skill MUST be executed by launching a subagent. Do NOT run synthesis in the main context.
Get the synthesis prompt, model, and pre-extracted data:
uv run --no-project $HOME/.claude/scripts/load_memory.py --synthesis-prompt
model=<model>. Second line: prompt_file=<path>.Launch a synthesis subagent (foreground — manual /synthesize always blocks so user sees results).
IMPORTANT: Do NOT read the prompt file. Pass the file path to the subagent and let it read the file itself. This avoids the main agent regenerating ~50K tokens of prompt content.
Task(
subagent_type: "general-purpose",
model: <model from first line>,
prompt: "Read <prompt_file path> and follow the instructions in it exactly. Use only Write and Bash tools."
)
Report the subagent's summary to the user.
Actions: implement, improve, document, analyze
Decisions: design, tradeoff, scope
Learnings: gotcha, pitfall, pattern
Lessons: insight, tip, workaround
Items in long-term memory can be moved to ## Pinned section (protected from decay):
Use when user asks about project status, wants to rename/move a project, clean up stale data, or mentions orphaned folders in ~/.claude/
Use when user wants to view, modify, or reset memory system configuration, or check token usage across memory tiers
Use when user wants to load a project's long-term and short-term memory mid-session — typically when running in light mode and Claude lacks project-specific context
Use when user wants to audit long-term memory for stale, incorrect, outdated, or duplicate entries, or correct memory entries based on new information
Use when user asks about past work, decisions, discussions, or topics that might have relevant history beyond loaded context
Use when user wants to explicitly preserve something permanently in long-term memory's Pinned section