jq-recipes.md

jq Recipes for Track CLI JSON

Practical one-liners that sit on top of track status --json. Copy/paste as-is; keep the CLI minimal and let jq do the shaping.

Quick Picks

• All in-progress tracks (id + title):

  track status --json | jq '.tracks[] | select(.status=="in_progress") | {id, title, kind}'

• Active next steps (title + next_prompt):

  track status --json | jq '.tracks[] | select(.status=="in_progress") | {title, next: .next_prompt}'

• Blocked tracks with parents (good for escalation lists):

  track status --json | jq '.tracks[] | select(.status=="blocked") | {title, parent: .parent_id, why: .summary}'

• Tracks missing next prompts (find weak breadcrumbs):

  track status --json | jq '.tracks[] | select((.status!="done") and (.next_prompt|length==0)) | {id, title, status}'

• Files grouped by feature (parent = feature id):

  FEATURE="u1v2w3x4"; track status --json \
    | jq --arg f "$FEATURE" '.tracks[] | select(.parent_id==$f) | {title, files}'

• Superseded tracks with their replacements (assumes title pattern "(OLD"/"NEW") or summary reference):

  track status --json | jq '.tracks[] | select(.status=="superseded") | {title, note: .next_prompt}'

• Per-feature completion snapshot (counts by status under a feature):

  FEATURE="u1v2w3x4"; track status --json \
    | jq --arg f "$FEATURE" '[.tracks[] | select(.parent_id==$f)] | group_by(.status) | map({status: .[0].status, count: length})'

• Stale in-progress work (needs timestamps in data):

  CUTOFF="2025-01-15T00:00:00Z"; track status --json \
    | jq --arg cutoff "$CUTOFF" '.tracks[] | select(.status=="in_progress" and .updated_at < $cutoff) | {title, updated_at}'

  If timestamps aren’t stored, skip this recipe.

Why jq First?

• Zero new flags, zero surface area.
• Encourages power use of existing JSON API.
• Serves as a proving ground; commands that get reused can later “graduate” into CLI flags.