chrislas/hyprvibe/beads
Claude Code and agent skill by chrislas/hyprvibe
- community
Checking that the SKILL.md file is still in the repository.
Distributed git-backed graph issue tracker for AI agents. Use this skill when working with Beads (bd) for tracking issues, tasks, decisions, and progress across multiple LLM sessions. Triggers include: setup/configuration of Beads, creating or managing issues/tasks/epics, session handoffs, tracking decisions, or any request to use Beads for project management.
Install
npx vskill@latest install chrislas/hyprvibe/beadsSKILL.md
Beads Operator
Beads (bd) is a distributed, git-backed, graph issue tracker designed for AI agents. This skill guides you to be an excellent Beads operator, ensuring work is properly tracked across sessions and machines.
Core Philosophy
Beads is the source of truth — not chat history.
Think in Beads, not chat. Translate all work into Beads nodes so another LLM can resume with near-zero context. Assume another LLM will pick this up later with no memory of this conversation.
Installation & Setup
First-time Setup
Install the Beads plugin for OpenCode:
On first use in a repository, run:
Daily Discipline
Run bd doctor once per day to verify health of the Beads system.
Operating Mode
1. Atomic, Graph-First Workflow
- One intent per node
- Tasks should be small and verifiable
- Track relationships explicitly:
depends_on,blocks,relates_to - Track Decisions separately from Tasks (immutable once done)
2. Git Discipline (Mandatory)
- Beads and git must stay in sync
- Never leave work unpushed
- Work is NOT complete until
git pushsucceeds
3. Always Ask
- "What changed?"
- "What is the next smallest verifiable step?"
Node Types & Requirements
Every Beads node MUST include:
- Purpose - Why this node exists
- Acceptance Criteria - How to know it's done
- Current State - Where things stand now
- Next Action - What to do next
Node Types
- Epic → Owns multiple Issues (large initiatives)
- Issue → Owns multiple Tasks (features or problems)
- Task → The only thing that goes Done quickly (atomic work)
- Decision → Rationale + alternatives rejected (immutable once Done)
- Handoff → Session-to-session continuity (for LLM handoffs)
Standard Tags
Use consistently across all nodes:
scope:<area>- infra, app, docs, monitoring, etc.host:<name>- custodian, ha, vps, etc.session:<name>- primary, remote, followup, etc.change:<type>- code, config, docsrisk:<low|med|high>- Risk levelprio:<p0|p1|p2>- Priority level
Quick Command Reference
Standard Response Format
For EVERY response when working with Beads, structure output as:
A) Beads Updates
B) Commands
Provide exact bd and git commands:
C) Working Notes
Terse notes on:
- Assumptions
- Risks
- Unknowns
D) Handoff Snapshot
Compact "resume here" paragraph for the next LLM session.
Session Workflow
Session Start
Before doing any work, gather:
- Repo path + git remote
- Objective for this session
- Other machines or LLM sessions that need handoff
Then propose:
- 1 Epic
- 2-5 Issues
- 5-12 Tasks
With explicit dependencies.
During Work
- Create nodes BEFORE starting work
- Update status as work progresses
- Track all decisions in Decision nodes
- Keep Tasks atomic and verifiable
- Sync frequently
Landing the Plane (MANDATORY)
When ending ANY work session, complete ALL steps:
- File Beads issues for remaining work
- Don't leave untracked work
- Run quality gates (if code/config changed)
- Tests, linters, etc.
- Update Beads status
- Close finished nodes
- Update in-progress nodes
- Sync + Push (CRITICAL)
``bash git pull --rebase bd sync git push git status # Must show clean + up to date ``
- Clean up local state
- Remove stashes
- Delete temp branches
- Verify nothing uncommitted
``bash git status ``
- Update/Create Handoff node
- Summary for next LLM
- Current state
- Next actions
Critical Rules for Session End
- Work is NOT complete until
git pushsucceeds - NEVER stop with local-only changes
- NEVER say "ready to push" — YOU push
- If push fails, resolve and retry until it succeeds
Best Practices
Node Granularity
- Too Big: "Implement authentication system"
- Just Right: "Add login endpoint", "Add JWT validation middleware", "Add password hashing"
Decision Tracking
When making architectural or implementation decisions:
Dependencies
Always track dependencies explicitly:
Session Continuity
For multi-session work, always create/update a Handoff node:
Common Patterns
Starting a New Feature
Making a Decision
Ending a Session
Troubleshooting
Beads Not Syncing
Lost Context
Check the most recent Handoff node:
Conflicts
Remember
- Beads is the memory - Chat history is ephemeral
- Small tasks - Break down until tasks are < 1 hour
- Always push - Local work doesn't exist to other LLMs
- Track decisions - Future you will thank present you
- Handoffs matter - Next LLM session depends on good handoffs
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingopencode/beads-skill/beads/SKILL.md
Popularity
- Repository stars
- 143
- Repository forks
- 17
- 7-day trend score
- 25
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 19 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/chrislas/hyprvibe/beads)