This section documents version-specific API changes in @storybloq/storybloq — prioritise recent major/minor releases.
-
NEW: storybloq init --type orchestrator — initialise a multi-repo orchestrator project source
-
NEW: storybloq node add <name> — add a node to an orchestrator with --path, --stack, --role, --depends-on, --link flags source
-
NEW: storybloq node update <name> — modify node role and integration links source
-
NEW: storybloq node remove <name> — deregister a node from the orchestrator source
-
NEW: storybloq node list — enumerate all registered federation nodes source
-
NEW: --node <name> flag on storybloq ticket, storybloq phase, storybloq issue commands — operate on a federated node's .story/ from the orchestrator directory source
-
NEW: --cross-node-blocked-by <refs> flag on storybloq ticket update — gate orchestrator milestones on node ticket completion (e.g. engine:T-020,client:T-012) source
-
NEW: storybloq_node_add, storybloq_node_list, storybloq_node_update MCP tools — CLI node operations via agent API source
-
NEW: storybloq_node_init MCP tool — bootstrap a node's .story/ directory from the orchestrator context source
-
CHANGED: storybloq_recommend — now synthesises suggestions from federated node graphs instead of single projects source
-
NEW: /story federation skill — guided setup flow for multi-repo orchestration source
-
NEW: storybloq dispatch command — dispatch tickets and issues to parallel Agent View background sessions with --recommend, --all, --yes flags source
-
NEW: Branch awareness at PICK_TICKET — autonomous guide detects ticket/issue IDs in git branch names and blocks mismatched work source
-
NEW: branchStrategy: "per-ticket" config option — auto-create story/T-XXX-slug branches at PICK_TICKET source
-
NEW: storybloq ticket meta get|set|unset <id> <key> [<value>] — query and modify custom metadata on tickets source
-
NEW: storybloq issue meta get|set|unset <id> <key> [<value>] — query and modify custom metadata on issues source
-
NEW: ticket_meta_get, ticket_meta_set, ticket_meta_unset MCP tools — metadata operations on tickets via agent API source
-
NEW: issue_meta_get, issue_meta_set, issue_meta_unset MCP tools — metadata operations on issues via agent API source
-
NEW: skip_ticket completedAction value — release ticket/issue claims and route to HANDOVER with context (replaces manual state manipulation) source
-
CHANGED: storybloq status output — now includes configured review backends and unconfigured setup guidance source
-
Create new handover files exclusively; never modify or overwrite existing ones — handovers are append-only historical records that form the crash-safe recovery boundary source
-
Run storybloq snapshot before writing a handover so the next session's recap can show what changed since the last snapshot source
-
When working on a ticket and encountering an out-of-scope bug or improvement, create an issue immediately via storybloq issue create with severity and impact description rather than fixing it inline — this keeps the tracker growing organically and ensures discoveries are not lost source
-
Distinguish tickets (planned work) from issues (discovered problems): when unsure, default to creating an issue — it can be promoted to a ticket later without data loss source
-
Check existing lessons via storybloq_lesson_digest before creating new lessons to avoid duplication — handovers are narrative, lessons are structured, tagged, and ranked source
-
In review findings disposition, use deferred only when genuinely wanting a new issue created (it auto-files), and reserve contested for genuine false positives, not valid-but-deferred work source
-
Use targeted autonomous mode (/story auto T-XXX ISS-YYY ...) to work through specific items in order and skip the branch-affinity check, avoiding unrelated commits contaminating feature branches source
-
Gate every irreversible action (push, deploy) on the byte-review verdict, never the implementer's word — byte-review agents verify reports against repo reality (commits exist, RED→GREEN re-derived, test counts logged) source
-
Enrich items in orchestrator mode using the template (CONTEXT / VERIFIED STATE @ sha / SCOPE / OUT OF SCOPE / ACCEPTANCE / PITFALLS / VERIFICATION / SIZING) with byte-verified facts so weaker models can trust the spec source
-
In orchestrator mode, maintain one pen per repo — never run two waves or edit while a wave executes in the same repo, as read-write conflicts bypass review gates source
-
Batch ships by collecting ~3 review-clean items per push (one deploy per ship), and ship risky/L-sized items alone fenced by pushes before and after source
-
Batch agent-discovered follow-ups for orchestrator filing (agents report via followUps structured output, the orchestrator files after checking existing items) rather than letting agents file issues — this prevents duplicate tracking and keeps the pen authority in the orchestrator source
-
When configuring storybloq via storybloq config set-overrides --json, pass ONLY the recipeOverrides object (e.g. {"maxTicketsPerSession": 10}), never top-level fields like version or project source
-
Never skip the active-session guard by auto-selecting a ticket — always surface via AskUserQuestion when an active session exists, and require explicit user choice (Resume / Cancel / Monitor) for each session source
