andresharpe/dotbot/status

Project Status Dashboard

Gather data from all available sources and present a structured status dashboard.

Data Collection

Collect the following in parallel where possible:

1. Task Queue (MCP tools)

• Call task_list to get all tasks with their status, priority, and assignments
• Call task_get_next to get the next actionable task
• Compute counts by status: todo, analysing, analysed, in-progress, done, needs-input, skipped

2. Session State (MCP tool)

• Call session_get_state to get current session info (if an active session exists)

3. Git Status (shell)

• Run git status --short to check working tree cleanliness
• Run git log --oneline -5 to show recent commits
• Run git branch --show-current to identify the current branch

4. Verification Gates (shell)

• Run pwsh .bot/hooks/verify/00-privacy-scan.ps1 — privacy/secrets scan
• Run pwsh .bot/hooks/verify/01-git-clean.ps1 — working tree cleanliness

Presentation Format

Present results as a structured dashboard:

## Task Queue
| Status       | Count |
|--------------|-------|
| todo         | N     |
| analysed     | N     |
| in-progress  | N     |
| done         | N     |
| needs-input  | N     |
 
## Next Up
<Highest priority task in `todo` or `analysed` status — show ID, title, priority>
 
## Currently Active
<Any tasks in `analysing` or `in-progress` status — show ID, title, assignee>
 
## Blocked / Needs Input
<Tasks in `needs-input` status — show ID, title, reason>
 
## Session
<Active session info, or "No active session">
 
## Git Status
Branch: <branch>
Clean: <yes/no>
Recent commits:
  <last 5 commits>
 
## Verification
- Privacy scan: PASS/FAIL
- Git clean: PASS/FAIL
 
## Suggestions
<Actionable next steps based on the current state, e.g.:
 - "3 tasks are analysed and ready for implementation"
 - "1 task needs input — review task T-XXXXXXXX"
 - "Working tree has uncommitted changes — commit or stash before starting next task">

Rules

• Read-only: Do not modify any state, tasks, or files
• Show all statuses: Include zero-count rows so the full picture is visible
• Be honest: If an MCP tool or hook fails, report the error rather than hiding it
• Actionable suggestions: Always end with 1-3 concrete next steps based on what the data shows