Run an ad-hoc chain of pm-skills in the order given, passing the shared context to every step.
Parse the input
$ARGUMENTS carries a chain expression followed by free-form context.
- Extract flags first, wherever they appear:
--auto,--force-auto,--dry-run,--thread. - The chain expression is exactly the leading separator-joined list of skill names (
,and->equivalent; mixing allowed). It ends after the first name that is NOT followed by a separator; everything after that is the context, even when the context words look like skill names. A trailing separator with nothing after it is an error to surface, not to guess around. - Everything after the chain expression is the context. This command parses the BOUNDARY only; all skill-name validation belongs to the engine.
Examples: /chain define-problem-statement -> define-hypothesis --thread Mobile checkout drop-off runs two steps with --thread, context "Mobile checkout drop-off". Lowercase context is fine: /chain deliver-prd, deliver-user-stories mobile checkout redesign is two steps with context "mobile checkout redesign" (the boundary is the missing separator after deliver-user-stories, not the letter case).
Confirm and dispatch
- Restate in one line: the parsed steps in order, the flags, and the context.
- Invoke the
utility-pm-workflow-orchestratorskill fromskills/utility-pm-workflow-orchestrator/SKILL.mdwith that chain, context, and flags (a Mode B run). The engine validates every name pre-flight and owns all run rules; this command adds none. The grammar lives inskills/utility-pm-workflow-orchestrator/references/PARSE-CONTRACT.md(Mode B Chain Expression Contract). - Relay the engine's per-step blocks and terminal output, including the promotion suggestion for reusable chains.
Nothing is persisted beyond the engine's own gitignored run artifacts. To make a chain durable, follow the completion suggestion to utility-pm-workflow-builder. While the native engine path is EXPERIMENTAL, recommend --dry-run first.
Context from user: $ARGUMENTS