From speckit-claude
Show the migration backlog status — what NEW-NNN entries are built, deferred, retracted, or planned, plus next-actionable items. Reads scripts/lib/migration-backlog-data.ts, the machine-readable registry that mirrors the §13 inventory tables in the architecture spec.
How this skill is triggered — by the user, by Claude, or both
Slash command
/speckit-claude:migration-backlogThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Any of these phrases should trigger this skill when spoken in natural language — please invoke `/migration-backlog` when the user asks:
/migration-backlog)Any of these phrases should trigger this skill when spoken in natural language — please invoke /migration-backlog when the user asks:
$ARGUMENTS
Arguments map to the underlying CLI subcommands of scripts/migration-backlog.ts:
summary → default summary with next-actionable + deferred tableslist [status] → full list, optionally filtered (planned, built, deferred, retracted)show <tag> → details for a single entry (e.g., show NEW-047)next → only the planned-entries-with-no-blockers subsetRun the reporter script:
npx tsx scripts/migration-backlog.ts $ARGUMENTS
Output the reporter's markdown verbatim to the user. Do NOT summarize, paraphrase, strip emoji, or cut sections. The reporter produces a complete report and the user wants to see all of it. Only add a brief sentence before or after the report if you're highlighting something specific (e.g., "Here's the current migration backlog:" or "NEW-065 is blocked by NEW-079.").
If the user asks follow-up questions about a specific tag's dependencies or history, use show <tag> on that tag — don't re-run the full summary.
The registry at scripts/lib/migration-backlog-data.ts is the authoritative source for:
built / built (schema) / built (reduced) / deferred / retracted / plannedB3, P34.3, Workshop 3 / §3.12.6)A companion script scripts/validate-migration-backlog.ts runs in CI (via .github/workflows/migration-backlog-lint.yml) and fails the build on registry drift:
built* entryfile: value in the registry must exist on diskbuilt* entries must have both file and pr set (unless they use implemented_by to point at a primary entry)planned / deferred / retracted entries must have file: null and pr: nullThis means: you cannot ship a migration without updating scripts/lib/migration-backlog-data.ts — the CI gate blocks the PR.
If your project maintains narrative documentation tables over the migration data, when they drift from the registry, the registry wins — update the docs to match, not the other way around.
/build-orderThe /build-order scanner reads feature specs' **Required Migrations**: NEW-NNN frontmatter line and cross-references it against this registry. A feature spec with an unbuilt migration dep gets classified as blocked and shows the blocking tag in its "Waiting on" column — the scanner merges F## feature blockers and NEW-NNN migration blockers into one list.
The handshake is direct: scripts/build-order/scan.ts imports MIGRATION_BACKLOG directly from scripts/lib/migration-backlog-data.ts and resolves alias entries (those with implemented_by) inline at startup. No TSV intermediary is used. scripts/migration-backlog-dump.ts is a separate utility that emits TSV for external/bash consumers — it is not called by the build-order scanner.
This means a dev running /build-order at the start of their day gets a unified view: "these features are ready; these are blocked by an upstream feature; these are blocked by an unbuilt migration." See .claude/skills/build-order/SKILL.md "How dependencies are resolved" for the matching side of the handshake.
show <tag> to check its statusnpx tsx scripts/migrate.ts so you know what new tables / columns landed/build-order for that (different concept, different registry)schema_migrations directly for that.npx tsx scripts/validate-migration-backlog.ts) instead, which checks drift; the reporter is a human-readable summary not a lint.tsc catches typos in tags, statuses, and required fields at edit time — mistakes surface before they hit CI.npx claudepluginhub gratefuljinx77/speckit-claude --plugin speckit-claudeCreates structured, bite-sized implementation plans from specs or requirements before writing code. Useful for breaking down multi-step tasks into testable steps with file structure and task boundaries.