beads-priority-assignment

Use this skill whenever you bd create a issue or re-triage priority (bd update) so the queue reflects real severity. Follow bd doctor / your Beads version for the exact --priority values your CLI accepts (integers 0–4, labels like P0–P4, and/or names like critical / high / medium / low).

Harness mapping (dashboard + types/beads.ts)

Exports and the dashboard normalize numeric priority roughly as:

Tier	Typical bd inputs	Meaning in this harness
Critical	0, P0, critical	Drop other work: production down, active exploit, data loss, irreversible customer harm
High	1, P1, high	Urgent: broken main path, release blocker, failing CI on default branch, severe bug with weak workaround
Medium	2, P2, medium	Default scheduled work: most features, typical bugs, refactors with agreed dates
Low	3–4, P3–P4, low	Backlog: polish, nice-to-have, cleanup, research spikes with no near deadline

Values 3 and above both map to low in the static dashboard export — use them to order work inside the backlog if your bd install preserves numeric order.

Rubric (three axes)

Score mentally; the highest axis that applies usually wins (then downgrade one step if mitigations exist).

1. Urgency — Must this start today / this week / this sprint / later?
2. Impact — How many users or systems are affected, and how badly?
3. Risk if deferred — Does waiting increase security exposure, compliance exposure, cost, or cascade blockers?

Escalate one level if the issue blocks other ready work or sits on a critical path to a dated epic.

De-escalate one level if there is a documented workaround, the area is feature-flagged off, or the report is duplicate / unclear until clarified.

Defaults

• If the user states “P0” / “urgent” / “critical”, map to that tier — do not normalize down.
• If you have no signal after the rubric, use P2 / 2 / medium (harness default), not empty priority.
• Never inflate priority to look important; never leave priority unset on purpose when the CLI supports it.

Examples

Situation	Suggested priority
Auth service returning 500 for all logins	0 / P0 / critical
CI red on master, no merge until fixed	1 / P1 / high
New feature from roadmap, normal sprint	2 / P2 / medium
Typo in internal-only docs	3 or 4 / low
/triage auto-filed “CI failing on PR #n” (matches existing harness command)	1 / P1 / high

Commands (non-interactive)

bd create --type task --title "[<repo>] ..." --priority <value>
bd update <id> --priority <value>

After changing priority on an existing issue, add a short Beads comment: worklog: priority set to <value> because <one line>.