Issues built for
agents, not meetings.
Every issue tracker today was built for humans to communicate with each other. Shipchart's Worklist is built for AI agents to execute. Structured acceptance criteria, file scope, and one-click delegation.
An execution packet,
not a to-do item.
Every work unit contains everything an AI agent needs to go from zero to pull request in one shot. No comments to parse, no ambiguity to interpret. Title, description, acceptance criteria, file scope, and human notes — all structured data.
- ›Structured acceptance criteria with checkable items
- ›File scope tells agents exactly where to look
- ›Human notes for course correction between attempts
- ›Agent reports capture decisions, files changed, and issues
- ›Size classification: single pass, multi-step, or spike
Every issue is
one click from a PR.
The primary action on every work unit is Delegate. Hit the button, pick your agent, and the work unit's description, acceptance criteria, scope, and human notes are assembled into the prompt automatically. Status syncs live: delegated, in review, done.
- ›One-click delegation to Codex, Copilot, or Claude Code
- ›Prompt auto-assembled from structured work unit fields
- ›Status syncs automatically: delegated → in review → done
- ›Re-delegate with human notes for course correction
- ›Every delegation creates a backing work unit — no orphans
Work comes in
from everywhere.
Work units aren't just manually created. They appear from roadmap decomposition, error detection, Deep Plan output, ad-hoc delegations, AI conversations, and MCP tools. Every source feeds into one unified list. Every item is instantly delegatable.
- ›Source tracking shows where each work unit came from
- ›Errors from Sentry become work units with full context
- ›Deep Plan explores your codebase and pre-fills scope and criteria
- ›MCP tools let Claude, Cursor, or any AI assistant create tasks
- ›Ad-hoc delegations always create a backing work unit
One list. Every
project. Both sources.
Native work units and external issues from Linear or GitHub appear in one unified view. Filter by status, size, source, or project. Portfolio-wide or scoped to a single project. Native items load instantly; external issues stream in alongside.
- ›Runs alongside Linear and GitHub Issues, not instead of them
- ›Portfolio-wide view across all projects
- ›Filter by status, size, source, or priority
- ›Native items load instantly from the database
- ›Available on web, iOS, Mac, CLI, and via MCP
Track work the way
agents think.
Structured issues, one-click delegation, and automatic status sync. The issue tracker for the agent-assisted era.