Slash Commands
Five commands for use inside Claude Code interactive sessions.
Slash commands are defined in .claude/commands/. Type them in a Claude Code session. They load the corresponding skill and run the workflow.
The commands
/source
Find cohorts, institutions, or sourcing paths matching a scientific or sourcing question. Runs the vcro-os query workflow against the local wiki.
CLI equivalent: vcro query "your question"
/bounty
Procurement workflow. Maps a procurement chain (specimen source to data delivery) with per-link evidence states. Budget and target N are optional. Runs vcro-bounty.
CLI equivalent: vcro bounty "your goal"
/onboard
Supply-side onboarding. Given an institution (name, URL, or PMC IDs), produces catalog, compliance, and pricing drafts. Surfaces gaps for institutional review.
CLI equivalent: vcro onboard "institution name"
/lint
Scan the wiki for gaps, drift, staleness, and broken links. Runs the 4 lint skills (gaps, consistency, staleness, connections) and aggregates findings.
CLI equivalent: vcro lint
/compile
Operator workflow. Extract, resolve, merge PMC papers into the wiki with parallel fan-out on extract. One Sonnet subagent per paper.
CLI equivalent: vcro compile PMC...
How they work
Each slash command maps to a skill file under .claude/commands/. When you type /source in Claude Code, the harness loads the skill instructions and the orchestrator follows the workflow defined there.
The slash commands and CLI commands run the same workflows. The difference is the entry point: slash commands run inside an interactive session, CLI commands run as a single-shot invocation.
Using in a session
# Launch an interactive session
vcro
# Inside the session, type:
/source AD plasma metabolomics cohorts longitudinal n>=200
# Or:
/compile PMC10103184 PMC9876543
# Or just ask naturally -- vcro-os routes the request without a slash command:
find me NSCLC FFPE blocks with paired RNA-seq
Slash commands are shortcuts. You can also ask questions in natural language and the orchestrator will route to the right workflow based on the verbs and nouns in the request.