← ClaudeAtlas

deferredlisted

Composable deferred-item operations. Called by other skills when a finding is deferred. Directly invocable as /deferred to view the current log.
rterakedis/leanwheel-skills · ★ 2 · AI & Automation · score 65
Install: claude install-skill rterakedis/leanwheel-skills
# Deferred Items (Composable) Composable operations for deferring findings → auto-scheduling as stories. Other skills call LOG-AND-SCHEDULE; user never manages queue. Directly invocable: `/deferred` — show log and linked story status. --- ## `docs/deferred-items.md` — the log Append-only single source of truth. Format: ```markdown | ID | Title | Detail | Source | Scheduled As | Date | |----|-------|--------|--------|--------------|------| | D-1 | {title} | {issue} | `{file}` | Story {epic}.{N} (slotted as AC) | {date} | | D-2 | {title} | {issue} | `{file}` | Story {epic}.{N} (new remediation) | {date} | ``` --- ## LOG-AND-SCHEDULE Input: `title`, `detail`, `source` (where issue found). Called by: code-review/correct-course/investigate/quick-dev when triaging defer findings. **Named trigger — observed test flake or hang.** A test suite that flakes or hangs is logged here the moment it is observed, never waved off as "passes in isolation." A festering hang blocks the next epic-boundary gate and masquerades as agent failures. **Gate — surface a decision instead of logging it.** Before completing the steps below: if the item you are about to write contains "decide: (a)/(b)/(c)", "needs a product decision", or similar, you have already done the analysis and the only thing being deferred is *asking*. Ask in that turn instead. A decision needs the owner when **any** of: both choices destroy or strand something the user created; the options differ in what they promise the