/context
Create, refresh, and manage your project context brief for use in web AI tools like ChatGPT, Perplexity, Gemini, and Claude web.
Overview
The /context command family helps you maintain and export project context for web-based AI tools. This is useful when you need to:
- Generate a comprehensive context brief for pasting into web AI tools
- Export a concise excerpt for quick context loading
- Record important decisions and discoveries in your context file
- Keep external AI tools informed about your project status
Subcommands
The /context command has three subcommands, each with a specific purpose:
/agileflow:context:full
Generate or refresh the comprehensive project context brief.
Creates or updates docs/context.md with managed sections containing:
- What we're building and current focus
- Feature map and tech/tooling summary
- Key decisions (ADRs) and architecture snapshot
- Testing & CI status
- Recent progress
- Risks and next steps
This file is designed to be pasted into ChatGPT, Perplexity, Gemini, or Claude web to give external AI tools full project context.
Usage:
/agileflow:context:fullBehavior:
- Reads existing
docs/context.mdif present - Gathers data from status.json, ADRs, epics, stories, and CHANGELOG
- Updates only managed sections (between
<!-- MANAGED_SECTION -->markers) - Preserves any user-written content outside managed sections
- Shows diff before applying changes
- Requires confirmation before writing
/agileflow:context:export
Export a concise context excerpt for pasting into web AI tools.
Reads docs/context.md and outputs a condensed excerpt (≤300 lines) ready to copy and paste into ChatGPT, Perplexity, Gemini, or Claude web.
This is a read-only command - no files are written.
Usage:
/agileflow:context:exportBehavior:
- Reads
docs/context.md - Extracts key sections (overview, current focus, tech stack, decisions, feature map, next steps)
- Outputs concise excerpt in a code block
- Includes instructions for pasting into web AI
- Does not require confirmation
/agileflow:context:note
Append a timestamped note to the context file.
Quickly add a note to docs/context.md under the "Notes" section. Useful for recording decisions, observations, or reminders that should persist in project context.
Usage:
/agileflow:context:note NOTE="Your note here"Arguments:
| Argument | Required | Description |
|---|---|---|
| NOTE | Yes | 1-5 line note text to append |
Behavior:
- Reads or creates
docs/context.md - Creates a "Notes" section if it doesn't exist
- Appends note with ISO timestamp (e.g.,
2024-12-31T10:30:00Z) - Shows diff before applying changes
- Requires confirmation before writing
Examples
Generate Full Context
/agileflow:context:fullCreates or updates docs/context.md with comprehensive project information:
# Project Context Brief
**Last Updated**: 2025-12-24
**Project**: AgileFlow - AI-powered project management
## What We're Building
AgileFlow is a comprehensive project management system powered by AI agents. It provides automated task management, documentation synchronization, and intelligent context preservation.
## Current Focus
- Implementing context compacting for long conversations
- Improving research workflow integration
- Expanding agent capabilities
## Feature Map
- **Core**: Project setup, status tracking, agent coordination
- **Documentation**: Automatic sync, README generation, ADR management
- **Agents**: 149 specialized agents for different tasks
- **Research**: Structured research notes and integration
## Tech Stack
- **Language**: TypeScript
- **Framework**: Next.js 16 (App Router)
- **UI**: React with Tailwind CSS
- **Package Manager**: npm with legacy-peer-deps
- **Deployment**: Vercel
## Key Decisions
- ADR-0001: Use TypeScript for type safety
- ADR-0002: Use React component library
- ADR-0003: Next.js for web framework
## Architecture Snapshot
- Monorepo structure with packages/ and apps/
- CLI package with 155 commands
- Website and documentation sites
- Agent-based command coordination
## Testing & CI
- Unit tests: Jest
- E2E tests: Playwright
- CI/CD: GitHub Actions
- Pre-commit hooks for validation
## Recent Progress
- Completed context compacting feature
- Improved documentation sync workflow
- Added 5 new specialized agents
## Risks
- Token limits with large projects
- Documentation drift during rapid development
## Next Steps
- Implement automated test coverage reporting
- Add real-time collaboration features
- Expand agent capabilities for code generationExport for Web AI Tools
/agileflow:context:exportOutputs a concise excerpt (≤300 lines) ready to paste:
# Project Context Brief
Last updated: 2025-12-24
Project: AgileFlow - AI-powered project management
## What We're Building
AgileFlow provides automated task management, documentation, and AI agents.
## Current Focus
- Context compacting for conversations
- Research workflow integration
- Agent capability expansion
## Tech Stack
- TypeScript, Next.js 16, React, Tailwind CSS, Vercel
## Key Decisions
- ADR-0001: Use TypeScript
- ADR-0002: Use React components
- ADR-0003: Use Next.js
## Feature Map
- Core: Setup, tracking, coordination
- Documentation: Sync, README, ADRs
- Agents: 149 specialized agents
- Research: Structured notes
## Next Steps
- Automated test coverage
- Real-time collaboration
- More agent capabilities
---
Paste this excerpt into your web AI tool (ChatGPT, Perplexity, Gemini, Claude web) to load project context.Add a Quick Note
/agileflow:context:note NOTE="User reported critical bug with OAuth flow on mobile"Appends a timestamped note to the "Notes" section of docs/context.md:
## Notes
- **[2025-12-24T10:30:00Z]**: User reported critical bug with OAuth flow on mobile
- **[2025-12-23T14:15:00Z]**: Completed initial research on TypeScript 5.4 features
- **[2025-12-22T09:00:00Z]**: Decided to upgrade to React 19 in next sprintWorkflow Examples
Keep Context Fresh
Keep context updated for better AI assistance:
# After each major milestone
/agileflow:context:full
# Before starting work in web AI
/agileflow:context:export
# Copy output and paste into ChatGPT/PerplexityUse in Web AI Conversations
Before asking web AI to help with your project:
# Export concise context
/agileflow:context:export
# Copy output and paste as initial message in ChatGPT
# Then ask your question with full project contextCapture Important Discoveries
/agileflow:context:note NOTE="TypeScript 5.4 improves type inference for generics"
/agileflow:context:note NOTE="Need to upgrade Node.js to 20+ for new features"
/agileflow:context:note NOTE="Team prefers MUI components over shadcn/ui"Output Modes
/agileflow:context:full (Create/Update Context)
Input: Gathers information from:
status.json- Current work and story statusbus/log.jsonl- Recent progress messagesdocs/03-decisions/- Architecture decisions (ADRs)docs/04-architecture/- System design documentationdocs/05-epics/- Feature roadmap and epicsCHANGELOG.md- Release historypackage.json- Project metadata
Output: Updates docs/context.md with managed sections while preserving custom content
Key Behavior:
- Diff-first: Shows changes before writing
- Managed sections: Only updates content between
<!-- MANAGED_SECTION -->markers - Preserves user content: Never modifies content outside managed sections
- Requires confirmation: User must approve changes
/agileflow:context:export (Copy to Web AI)
Input: Reads existing docs/context.md
Output: Concise excerpt (≤300 lines) ready to paste into ChatGPT, Perplexity, Gemini, or Claude web
Use Case: Load project context into web AI tool to continue work in a fresh conversation
Key Behavior:
- Read-only: No files are modified
- Self-contained: Includes instructions for pasting
- Format-ready: Already formatted for easy copy-paste
/agileflow:context:note (Append Timestamped Note)
Input: NOTE argument with 1-5 line note
Output: Appends to "Notes" section with ISO timestamp
Use Case: Quick capture of important events, decisions, or discoveries
Key Behavior:
- Diff-first: Shows changes before writing
- Creates section if missing: Automatically creates "Notes" section if needed
- Chronological: Notes appear in order with newest at bottom
- Timestamped: Every note includes ISO 8601 timestamp
Best Practices
Regular Context Updates
Keep context fresh for better AI assistance:
- After major milestones: Run
/agileflow:context:fullto capture completed work - Before web AI sessions: Run
/agileflow:context:exportand paste into ChatGPT - During development: Use
/agileflow:context:noteto record decisions and discoveries
# Recommended workflow
/agileflow:context:full # Update full context
/agileflow:context:export # Get concise version
# Paste into web AI for context-aware helpCapture Important Discoveries
Record technical insights and decisions:
/agileflow:context:note NOTE="TypeScript 5.4 improves type inference - consider upgrading"
/agileflow:context:note NOTE="Stripe API changed pagination in v2.0 - update integration"
/agileflow:context:note NOTE="Team consensus: use Postgres over MongoDB for this project"Use in Web AI Conversations
Maximize effectiveness of web AI tools by providing context:
- Export concise context:
/agileflow:context:export - Copy the output
- Paste into ChatGPT, Perplexity, or Gemini as your first message
- Ask follow-up questions with full project context
This ensures the web AI understands your codebase, architecture, and current focus.
Manage Context File Size
For large projects, manage context.md file size:
- Use
/agileflow:context:exportto check current context size - Archive old notes if they grow beyond 100+ entries
- Create separate research files for deep dives
- Use
/agileflow:research:askfor complex research topics
Diff-First Workflow
All context commands that modify files follow a consistent workflow:
- Show Changes: Command displays proposed changes as a diff
- Confirm: You review changes and approve or reject
- Apply: Changes are written only after your confirmation
This ensures you never lose important information and always know what's being modified.
Related Commands
/adr- Create architecture decision records/research- Create and manage research notes/docs-sync- Synchronize documentation/readme-sync- Update README files
Notes
- Context files are stored in
docs/context.mdas part of your local project documentation - Managed sections use HTML comments as markers and should not be edited manually
- The export command respects the 300-line limit to fit in web AI tool context windows
- Notes are timestamped in ISO 8601 format for consistency and easy sorting
On This Page
/contextOverviewSubcommands/agileflow:context:full/agileflow:context:export/agileflow:context:noteExamplesGenerate Full ContextExport for Web AI ToolsAdd a Quick NoteWorkflow ExamplesKeep Context FreshUse in Web AI ConversationsCapture Important DiscoveriesOutput Modes/agileflow:context:full (Create/Update Context)/agileflow:context:export (Copy to Web AI)/agileflow:context:note (Append Timestamped Note)Best PracticesRegular Context UpdatesCapture Important DiscoveriesUse in Web AI ConversationsManage Context File SizeDiff-First WorkflowRelated CommandsNotes