kata

kata is a local-first issue tracker designed for workflows where AI coding agents create, claim, update, and close issues as part of their work, while human supervisors oversee progress via a terminal UI or browser.

Design Philosophy #

Issue state lives in SQLite under KATA_HOME; the git repository contains only a small, secret-free .kata.toml configuration file. This keeps issue history out of commit history and avoids requiring a hosted tracker (GitHub Issues, Linear, Jira) as a dependency for development work.

A single Go binary with no runtime dependencies handles the full lifecycle.

Key Capabilities #

• Agents create, claim, relate, and close issues with evidence attached
• Human TUI and browser UI for oversight and triage
• GitHub sync (bidirectional, described in the operations guide)
• Optional remote daemon mode with PostgreSQL backend for team sharing
• Federation for distributed deployments
• In-process HTTP service for embedding in Go applications

Relation to Other kenn-io Tools #

kata integrates with roborev for correlating code review findings with issue records. forge provides the broader maintainer console context (PRs, CI) within which kata issues live.

Relevance to Archival #

kata’s local SQLite store is a natural archival target: issue history, agent activity records, and evidence attachments can be snapshotted into a DataLad dataset alongside the code. The .kata.toml in the repository provides the linkage between issue state and the commit graph.

No established workflow for DataLad-native kata archival exists yet.

See Also #

• forge – local-first maintainer console for PRs and issues across forges
• git-bug – distributed issue tracking stored directly in git objects
• github-backup – full JSON export of GitHub Issues and PRs
• roborev – AI code review database, integrates with kata