Board, Code Browser & Stats

ait board

Open the kanban-style TUI board for visual task management.

ait board

Launches a Python-based terminal UI (built with Textual) that displays tasks in a kanban-style column layout. All arguments are forwarded to the Python board application.

For full usage documentation — including tutorials, keyboard shortcuts, how-to guides, and configuration — see the Kanban Board documentation.

Requirements:

• Python venv at ~/.aitask/venv/ with packages: textual, pyyaml, linkify-it-py
• Falls back to system python3 if venv not found (warns about missing packages)
• Checks terminal capabilities and warns on legacy terminals (e.g., WSL default console)

ait codebrowser

Open the code browser TUI for file exploration with task annotations.

ait codebrowser

Launches a Python-based terminal UI (built with Textual) that provides a file tree, syntax-highlighted code viewer, and annotation gutter showing which aitasks contributed to each section of code. All arguments are forwarded to the Python codebrowser application.

For full usage documentation — including tutorials, keyboard shortcuts, how-to guides, and annotation details — see the Code Browser documentation.

Requirements:

• Python venv at ~/.aitask/venv/ with packages: textual, pyyaml
• Falls back to system python3 if venv not found (warns about missing packages)
• Checks terminal capabilities and warns on legacy terminals (e.g., WSL default console)

ait stats

Display task completion statistics and trends.

ait stats                  # Basic stats (last 7 days)
ait stats -d 14            # Extended daily view
ait stats -v               # Verbose with task IDs
ait stats --csv            # Export to CSV
ait stats -w sun           # Week starts on Sunday
ait stats --backlog-weeks 26   # Longer backlog history
ait stats --csv-backlog        # Export the weekly backlog series

Statistics provided:

1. Summary — Total completions, 7-day and 30-day counts
2. Backlog level — Weekly count of open tasks per category, with follow-up and genuine-work subtotals, a TOTAL OPEN row, and its parent / child partition
3. Backlog net flow — Weekly arrivals minus departures per category, with ARRIVALS / DEPARTURES / NET summary rows, explaining why the level moves
4. Daily breakdown — Completions per day (with task IDs in verbose mode)
5. Day of week averages — This week counts + 30-day and all-time averages per weekday
6. Pipeline timing — Average time in the implement and review/merge phases, for gated tasks
7. Label weekly trends — Per-label completions for last 4 weeks
8. Label day-of-week — Per-label averages by day of week (last 30 days)
9. Task type trends — Parent/child and issue type (feature/bug/refactor) weekly trends
10. Label + type trends — Issue types by label, weekly for last 4 weeks
11. Code agent trends — Weekly completion split by code agent for last 4 weeks
12. LLM model trends — Weekly completion split by normalized LLM model for last 4 weeks
13. Verified model rankings — Model scores per skill, aggregated across providers

The two backlog sections cover the last 8 weeks by default; --backlog-weeks changes the horizon. In both tables the columns run chronologically with the current week last, and in the net-flow table that final column is a partial week (marked Now*, with the dates it covers given below the table).

Data sources: The completion sections scan archived parent tasks (aitasks/archived/t*_*.md), archived child tasks (aitasks/archived/t*/), and compressed archives (numbered _bN/oldM.tar.gz bundles). They use the completed_at field, falling back to updated_at for tasks with status: Done, and prefer the merge_approved / review_approved gate-ledger stamps where a task has them.

The two backlog sections need open tasks as well, so they scan your active tasks alongside the archive, and they use a different clock: completed_at, falling back to updated_at for tasks with status: Done or Completed — never the gate-ledger stamps, which on a live task mean “in flight”, not “finished”. The two clocks agree on whether a task completed and can disagree on which week it landed in, so a small number of tasks are bucketed one week apart between the backlog sections and the completion sections.

Backlog population rules: a task counts as open from its created_at week until the week it departs, so Postponed tasks count as open; a Done task that has not been archived yet counts as departed; Folded tasks are excluded entirely, as are tasks with no frontmatter, no created_at, or a future date. The report footnotes how many tasks were excluded and why.

CSV export format: --csv writes one row per completed task: date, day_of_week, week_offset, task_id, labels, issue_type, task_type, implemented_with, codeagent, llm_model, created_at, category. Open in LibreOffice Calc for custom charts and pivot tables.

Open tasks are not rows in that table, so the backlog level cannot be reproduced from it — created_at there supports lead-time analysis, not backlog. The backlog series has its own export: --csv-backlog writes week_ending, category, open, arrived, departed, net, one row per category per week, oldest week first, including zero cells so the grid is dense and plottable. Its category values are raw namespaced keys (type:feature, kind:manual_verification) rather than the display labels the report prints, so they are safe to join on, and it contains only real categories — no subtotal or TOTAL OPEN rows for a pivot to double-count. --backlog-weeks sets its horizon too.

Interactive charts:

For interactive terminal charts (daily completions, weekday averages, top labels, issue types, code agents, LLM models, backlog level and net flow), use ait stats-tui. The TUI is launched directly or switched into from any other aitasks TUI (j in the TUI switcher). It uses the plotext package, installed and version-pinned by ait setup.

Next: Explain Utilities