tk-context

iheanyi's avatarfrom iheanyi

Dump full Tasuku context for agent consumption. Use when user says /tk:context or asks for full context, project state, or needs to understand the entire task landscape.

⭐18stars🔀0forks📁View on GitHub🕐Updated Jan 10, 2026⬇️2downloads

When & Why to Use This Skill

The tk-context skill is a powerful context-retrieval tool for Claude agents, designed to provide a comprehensive snapshot of the Tasuku project state. By dumping the full task landscape—including active work, blockers, architectural decisions, and key learnings—it ensures the AI agent maintains perfect continuity and situational awareness throughout complex development or project management workflows.

Use Cases

  • Session Initialization: Use '/tk:context' at the start of a new chat to instantly bring the AI agent up to speed on the current project state and recent progress.
  • Strategic Alignment: Request a full context dump to help the agent understand the relationship between different tasks, ensuring that new suggestions align with previous architectural decisions.
  • Blocker Identification: Quickly extract and summarize all 'blocked' tasks and their associated impediments to facilitate rapid problem-solving and resource reallocation.
  • Knowledge Transfer: Retrieve a structured list of project-specific rules and learnings to ensure the agent adheres to established coding standards or operational guidelines.
nametk-context
descriptionDump full Tasuku context for agent consumption. Use when user says /tk:context or asks for full context, project state, or needs to understand the entire task landscape.

Full Context Dump

Provide complete project context including all tasks, learnings, decisions, and notes.

Instructions

  1. Use the tk_context MCP tool to get the full .tasuku.json state
  2. Present the information in a structured, scannable format
  3. Highlight key items:
    • Tasks currently in progress (what's being worked on)
    • Blocked tasks and their blockers
    • Recent decisions that affect current work
    • Relevant learnings/rules

Output Format

## Current Work
[List in_progress tasks with owner if set]

## Ready Queue
[List ready tasks sorted by priority]

## Blockers
[List blocked tasks and what's blocking them]

## Key Decisions
[Recent or relevant architectural decisions]

## Learnings & Rules
[Important insights, especially any marked as rules]

## Notes
[Any notes attached to active tasks]

Notes

  • This is meant for getting oriented at the start of a session
  • Focus on actionable information
  • Skip archived tasks unless specifically requested
  • If context is large, summarize done tasks as a count