prioritise

prioritise — rank the queue, dependency-aware

Answer "what should I do next?" by ranking the ready/ briefs. The score is transparent and generic: due urgency + importance + quick-win (low effort) + lead-time + a bonus for briefs that unblock others, plus any priority tag weights defined in tasks.toml.

Dependencies are real: a brief lists blockers: (brief ids or external text) and/or depends:<id> tags. A dependency counts as met only once its target brief is done, so the queue self-unblocks as work completes.

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prioritise.py"               # ranked, all of ready/
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prioritise.py" --top 5       # top 5
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prioritise.py" --actionable  # only what Claude can run now
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prioritise.py" --json        # machine-readable
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prioritise.py" --context cloud  # skip requires-local / requires-repo

The script resolves the queue via $CLAUDE_TASKS_DIR → nearest .tasks/ → ~/tasks.

How to use it

• "What should I do next?" → run it (optionally --top 5), then state the top few back in your own words: id, title, why it's ranked there (due / importance / unblocks N), and whether it's actionable now or blocked / needs them.
• "What can you pick up?" → --actionable (only autonomy: full, unblocked, runnable). This is also what an autonomous loop would call (--actionable --context cloud --json).
• "What's blocked?" → run it and read the ⛔blocked(<unmet deps>) flags; offer to action the blocker first (a brief that unblocks others is ranked higher for exactly this).

Flags in the output

• ⛔blocked(...) — not actionable; the listed dependencies aren't done yet.
• 🙋needs-input / 🙋blocked — waiting on the user, not auto-runnable.
• →N — this brief unblocks N others (clearing it has leverage).
• ⏱lead — the effort lead time means it must start now to hit its due date.

Defining your own priority tags

To weight specific tags (e.g. a "frog"/"urgent" convention), add to tasks.toml:

[priority]
tag_weights = { urgent = 500, frog = 400 }

Briefs carrying those tags get the weight added to their score. Empty by default — the base ranking stays purely mechanical (due / importance / effort / unlocks).