// search past sessions

eagle-mem search "auth middleware"

Search across session summaries, Claude memories, and indexed code using FTS5. Use this when you know you worked on something but can’t remember the details.

eagle-mem search --timeline

Shows recent sessions in chronological order — useful for catching up after a break.

// project overview

eagle-mem overview

View the overview that gets injected into every Claude Code session. This is what Claude reads first.

eagle-mem overview set "My app is a..."

Set a custom overview. Or let Claude write one: run /eagle-mem-overview inside a session.

// sync & maintain

eagle-mem refresh

Full re-sync: re-scans codebase, re-indexes source files, syncs new Claude Code memories and tasks. Run this after major changes (new packages, restructured dirs, large pull).

eagle-mem prune

Clean up old observations and orphaned code chunks. Safe to run anytime.

// more commands

• eagle-mem scan — re-analyze codebase structure
• eagle-mem index — re-index source files for code search
• eagle-mem memories — view or sync mirrored Claude Code memories and plans
• eagle-mem tasks — view mirrored Claude Code tasks
• eagle-mem update — re-deploy hooks after npm update
• eagle-mem uninstall — remove hooks and optionally delete data

// /eagle-mem-search

Search past sessions from within Claude Code. Claude interprets results and connects them to your current work — better than raw terminal search when you need context, not just a match.

// /eagle-mem-overview

Build a rich project briefing. Claude reads your README, entry points, and git history to write 2-3 paragraphs that capture what the project does. Injected at every session start.

// /eagle-mem-tasks

Break complex work into tasks that survive /compact. Uses Claude Code’s native TaskCreate/TaskUpdate. When context fills up, Eagle Mem re-injects task state so Claude picks up where it left off.

/eagle-mem-scan
Analyze codebase structure

/eagle-mem-index
Index source files for code search

/eagle-mem-memories
View & sync mirrored memories

/eagle-mem-prune
Database cleanup & hygiene

// automatic memory

Six lifecycle hooks fire on every session event. You never run a “save” command — Eagle Mem captures context automatically.

• SessionStart — injects overview, summaries, memories, plans, and in-progress tasks
• PreToolUse — rewrites noisy commands (learned rules), detects redundant reads
• UserPromptSubmit — FTS5 search for relevant past memories
• PostToolUse — records file touches, mirrors memories/plans/tasks, tracks modifications
• Stop — extracts <eagle-summary>, strips <private> tags
• SessionEnd — re-syncs task statuses, marks session completed

// project overview

Model-driven overviews — Claude reads your README, entry points, and recent git history to understand what the project does, not just its file counts.

• Auto-trigger — if no overview exists, SessionStart prompts Claude to build one
• Synthesized — purpose, current state, and key tech in 2-3 sentences
• Injected — overview surfaces at every session start
• Scan fallback — structural scan for repos with no README

// taskaware compact loop

Break complex multi-step work into tasks using Claude Code’s native TaskCreate/TaskUpdate with dependency support.

• Plan — create tasks with TaskCreate + addBlockedBy
• Execute — work on one task at a time
• Compact — run /compact when context fills up
• Resume — Eagle Mem re-injects mirrored task state + memory

// claude code mirror

Automatically mirrors Claude Code’s built-in memories, plans, and tasks into Eagle Mem’s SQLite + FTS5 — searchable and injected at session start.

• Memories — intercepts writes to ~/.claude/projects/*/memory/
• Plans — captures plan files from ~/.claude/plans/
• Tasks — mirrors TaskCreate/TaskUpdate + re-syncs at session end
• eagle-mem memories sync — backfill everything in one command

// privacy & safety

Everything stays local. One SQLite file, no network calls, no external services.

• <private> tags — stripped before any data reaches the database
• WAL mode — concurrent session safety across 4-5 instances
• Secret redaction — API keys and tokens scrubbed from observations
• Per-project isolation — each project gets its own memory bucket

// token savings

Eagle Mem actively reduces token consumption through learned behavior, not hardcoded rules.

• Command rewriting — PreToolUse rewrites noisy Bash commands to pipe through head -N via updatedInput. Rules are learned by the curator from real usage.
• Read-after-modify — detects when you read a file that was just edited or written, nudges that the diff is already in context
• Read dedup — files read 3+ times in a session get a soft nudge that contents are likely already in context
• Self-learning — the curator analyzes observation data to discover which commands are noisy and generates project-specific rules

// compact-safe reload

When Claude Code compacts your conversation, Eagle Mem detects it and re-injects the full context — overview, memories, tasks, and recent session history.

• Trigger-aware — shows startup vs compact vs clear vs resume
• Stale filtering — skips tasks older than 7 days and noisy auto-captured summaries
• Smart project ID — resolves via git root, handles monorepos correctly