e58d60b802
Add granular control over MCP tool response sizes to minimize context
window usage while maintaining full functionality when needed.
## Output Control Parameters
### Read Operations (ready, list, show, blocked)
- `brief`: Return BriefIssue {id, title, status, priority} (~97% smaller)
- `fields`: Custom field projection with validation - invalid fields raise ValueError
- `max_description_length`: Truncate descriptions to N chars with "..."
- `brief_deps`: Full issue but dependencies as BriefDep (~48% smaller)
### Write Operations (create, update, close, reopen)
- `brief`: Return OperationResult {id, action, message} (~97% smaller)
- Default is brief=True for minimal confirmations
## Token Savings
| Operation | Before | After (brief) | Reduction |
|-----------|--------|---------------|-----------|
| ready(limit=10) | ~20KB | ~600B | 97% |
| list(limit=20) | ~40KB | ~800B | 98% |
| show() | ~2KB | ~60B | 97% |
| show() with 5 deps | ~4.5KB | ~2.3KB | 48% |
| blocked() | ~10KB | ~240B | 98% |
| create() response | ~2KB | ~50B | 97% |
## New Models
- BriefIssue: Ultra-minimal issue (4 fields, ~60B)
- BriefDep: Compact dependency (5 fields, ~70B)
- OperationResult: Write confirmation (3 fields, ~50B)
- OperationAction: Literal["created", "updated", "closed", "reopened"]
## Best Practices
- Unified `brief` parameter naming across all operations
- `brief=True` always means "give me less data"
- Field validation with clear error messages listing valid fields
- All parameters optional with backwards-compatible defaults
- Progressive disclosure: scan cheap, detail on demand
---
## Filtering Parameters (aligns MCP with CLI)
Add missing filter parameters to `list` and `ready` tools that were
documented in MCP instructions but not implemented.
### ReadyWorkParams - New Fields
- `labels: list[str]` - AND filter: must have ALL specified labels
- `labels_any: list[str]` - OR filter: must have at least one
- `unassigned: bool` - Filter to only unassigned issues
- `sort_policy: str` - Sort by: hybrid (default), priority, oldest
### ListIssuesParams - New Fields
- `labels: list[str]` - AND filter: must have ALL specified labels
- `labels_any: list[str]` - OR filter: must have at least one
- `query: str` - Search in title (case-insensitive substring)
- `unassigned: bool` - Filter to only unassigned issues
### CLI Flag Mappings
| MCP Parameter | ready CLI Flag | list CLI Flag |
|---------------|----------------|---------------|
| labels | --label (repeated) | --label (repeated) |
| labels_any | --label-any (repeated) | --label-any (repeated) |
| query | N/A | --title |
| unassigned | --unassigned | --no-assignee |
| sort_policy | --sort | N/A |
---
## Documentation & Testing
### get_tool_info() Updates
- Added `brief`, `brief_deps`, `fields`, `max_description_length` to show tool
- Added `brief` parameter docs for create, update, close, reopen
- Added `brief`, `brief_deps` parameter docs for blocked
### Test Coverage (16 new tests)
- `test_create_brief_default` / `test_create_brief_false`
- `test_update_brief_default` / `test_update_brief_false`
- `test_close_brief_default` / `test_close_brief_false`
- `test_reopen_brief_default`
- `test_show_brief` / `test_show_fields_projection` / `test_show_fields_invalid`
- `test_show_max_description_length` / `test_show_brief_deps`
- `test_list_brief` / `test_ready_brief` / `test_blocked_brief`
---
## Backward Compatibility
All new parameters are optional with sensible defaults:
- `brief`: False for reads, True for writes
- `fields`, `max_description_length`: None (no filtering/truncation)
- `labels`, `labels_any`, `query`: None (no filtering)
- `unassigned`: False (include all)
- `sort_policy`: None (use default hybrid sort)
Existing MCP tool calls continue to work unchanged.
bd - Beads
Distributed, git-backed graph issue tracker for AI agents.
Beads provides a persistent, structured memory for coding agents. It replaces messy markdown plans with a dependency-aware graph, allowing agents to handle long-horizon tasks without losing context.
⚡ Quick Start
# Install (macOS/Linux)
curl -fsSL https://raw.githubusercontent.com/steveyegge/beads/main/scripts/install.sh | bash
# Initialize (Humans run this once)
bd init
# Tell your agent
echo "Use 'bd' for task tracking" >> AGENTS.md
🛠 Features
- Git as Database: Issues stored as JSONL in
.beads/. Versioned, branched, and merged like code. - Agent-Optimized: JSON output, dependency tracking, and auto-ready task detection.
- Zero Conflict: Hash-based IDs (
bd-a1b2) prevent merge collisions in multi-agent/multi-branch workflows. - Invisible Infrastructure: SQLite local cache for speed; background daemon for auto-sync.
- Compaction: Semantic "memory decay" summarizes old closed tasks to save context window.
📖 Essential Commands
| Command | Action |
|---|---|
bd ready |
List tasks with no open blockers. |
bd create "Title" -p 0 |
Create a P0 task. |
bd dep add <child> <parent> |
Link tasks (blocks, related, parent-child). |
bd show <id> |
View task details and audit trail. |
bd quickstart |
Interactive guide for AI agents. |
🔗 Hierarchy & Workflow
Beads supports hierarchical IDs for epics:
bd-a3f8(Epic)bd-a3f8.1(Task)bd-a3f8.1.1(Sub-task)
Stealth Mode: Run bd init --stealth to use Beads locally without committing files to the main repo. Perfect for personal use on shared projects.
📦 Installation
- npm:
npm install -g @beads/bd - Homebrew:
brew install steveyegge/beads/bd - Go:
go install github.com/steveyegge/beads/cmd/bd@latest
Requirements: Linux (glibc 2.32+), macOS, or Windows.
🌐 Community Tools
- beads_viewer - Keyboard-driven terminal UI with kanban board, insights panel, and graph view. Built by @Dicklesworthstone.
- beads.el - Emacs UI to browse, edit, and manage beads. Built by @ctietze.
- beads-ui - Local web interface with live updates and kanban board.
npx beads-ui start. Built by @mantoni. - bdui - Real-time terminal UI with tree view, dependency graph, and vim-style navigation. Built by @assimelha.
- perles - Terminal UI with BQL (Beads Query Language) and multi-view kanban. Built by @zjrosen.
- vscode-beads - VS Code extension with issues panel and daemon management. Built by @jdillon.