session-start

Session Start

Load critical context and recent history at the start of each session or after /clear.

Contract

Inputs:

• Project root
• Current date
• Existing .agents/memory/ and .agents/sessions/ files

Outputs:

• Loaded repo memory summary
• Today's session context (if any)
• Open next-steps surfaced from the latest session log

Creates/Modifies:

• No required writes
• May create today's session file only through session-documenter

External Side Effects:

• None

Confirmation Required:

• Before creating or rewriting session files when paths are ambiguous

Delegates To:

• session-documenter to track the active session
• rules-capture if startup reveals uncaptured preferences

Workflow

1. Read Repo Memory (CRITICAL — source of truth)

Read the repo's memory files at .agents/memory/. These are the source of truth for current project context: architecture, deployment state, migrations, known gotchas, and operating decisions. Apply the staleness and temporary-file checks defined in the global CLAUDE.md / MEMORY.md before citing any entry.

2. Read Today's Session File

Read today's session to understand what was already done before /clear:

Read today's session file at .agents/sessions/YYYY-MM-DD.md (where YYYY-MM-DD is today's date).

If the file exists, it shows tasks completed, decisions made, files changed, mistakes, and next steps. If it doesn't exist yet, this is a fresh session day.

3. Activate Session Documenter

Run the session-documenter skill to track work throughout the session.

4. Confirmation

After reading memory and the session file, provide a brief confirmation (5-7 bullet points max):

• Critical rules understood
• Repo memory loaded (flag any stale last_verified or status: temporary files)
• Today's session context loaded (if it exists)
• Session documenter active
• Open next-steps surfaced from the latest session log (if any)

Related Skills

• session-end — Document session before clearing context
• session-documenter — Tracks work throughout the session