session-start

Loads CLAUDE.md, memory files, and architecture context into a session brief at startup.

1|Updated Jun 27, 2026
One-click install
npx skills add https://github.com/vickysrawat/AI-Assisted-development --skill session-start-vickysrawat
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: session-start
Source: https://github.com/vickysrawat/AI-Assisted-development/tree/main/skills/session-start
Command: npx skills add https://github.com/vickysrawat/AI-Assisted-development --skill session-start-vickysrawat

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Starting a new Claude Code session usually means manually re-explaining the project stack, recent decisions, and active work areas. This Skill automates that warm-up by reading project context files in one pass and printing a concise session brief. ## Core Features & Use Cases - One-pass context loading: Reads CLAUDE.md, memory files, topic notes, and architecture docs, then prints a session brief under 20 lines. - Feature gate pre-loading: Holds the ICEA approval gate rule in active context so implementation code is never generated before an approved ICEA exists. - Session identity resolution: Best-effort population of .claude/session-context.json (AD UPN, git email, optional Azure DevOps PAT identity) so the governance audit trail can attribute events to a verified person. - Use Case: A developer opens a repo on Monday morning, runs /session-start, and immediately sees the stack, target branch, last decision, last fix, and active modules without typing any context manually. ## Quick Start Run the /session-start command at the beginning of your Claude Code session to load all project context and see the session brief.

Frequently Asked Questions about session-start

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I load project context automatically at the start of a Claude Code session?

Run the /session-start command, which reads CLAUDE.md, memory files, topic notes, and architecture docs in one pass and prints a session brief under 20 lines. It never scans source files and never asks you for information.

What files does session-start read to build the session brief?

It reads CLAUDE.md, memory/MEMORY.md, up to five recent memory/topic-*.md files, memory/dream-log.md, architecture.md, architecture-deployment.md, and the graph index. Missing files are skipped silently.

Does session-start work without a CLAUDE.md file?

No. If CLAUDE.md is missing entirely, it outputs a warning telling you to run /setup-init to set up the project, then stops. Other missing files are skipped silently.

How does session-start handle identity for the audit trail?

It best-effort populates .claude/session-context.json using the Windows AD UPN, git email, and optionally a PAT-verified Azure DevOps identity. The result is cached for 7 days, and if nothing resolves, attribution falls back to the OS login without blocking the session.

What is the feature gate that session-start pre-loads?

It holds a rule in context requiring an approved ICEA document before any implementation code is generated for new features. Bug fixes, refactoring, questions, and /skip-icea-prefixed requests are exempt.

Why does session-start update Last-cited stamps in topic files?

It stamps memory topic files whose content actually influenced the session brief, feeding citation telemetry to /dream-audit. Files read but not used are not stamped, which keeps staleness analysis meaningful.