CLI Reference

Use --help on any command for exact options:

npx jumpspace --help
npx jumpspace plan --help
npx jumpspace bootstrap --help

Setup

Command	Purpose
`init`	Create starter Jumpspace files.
`init --auto`	Discover docs and create a better starter config.
`init --ci github`	Add GitHub CI workflow support.
`add-skill --codex`	Add Codex guidance.
`add-skill --claude`	Add Claude Code guidance.
`release doctor`	Check package release readiness.
`release install-doctor`	Check which installed CLI is active.

Command	Purpose
`scan`	Parse Markdown task blocks and write `.jumpspace/index.json`.
`list`	List indexed tasks.
`find <query...>`	Search indexed tasks.
`context <id>`	Print task context.
`ask <question...>`	Return an evidence summary for a question.
`semantic build`	Build local semantic retrieval index.
`semantic status`	Show semantic index readiness.
`semantic search <query...>`	Search semantic task vectors.
`semantic eval`	Compare retrieval modes.

Command	Purpose
`bootstrap discover`	Detect docs and recommend config globs.
`bootstrap context <paths...>`	Export heading context for an AI proposer.
`bootstrap propose <paths...> --file <file>`	Create a proposal file.
`bootstrap validate --file <file>`	Validate a proposal.
`bootstrap apply --file <file>`	Apply proposed task blocks.
`bootstrap apply --dry-run`	Preview Markdown changes.

Command	Purpose
`plan review <id>`	Review plan state and issues.
`plan save <id> --file <file>`	Persist a durable plan.
`plan show <id>`	Show a task plan.
`plan validate <id>`	Validate plan structure and execution state.
`ready`	List approved tasks ready for execution.
`execute <id>`	Print an execution packet.
`work <id>`	Print the main agent start packet.
`next <id>`	Show pending unblocked plan steps.
`step complete <id> <step-id>`	Mark a plan step complete with evidence.
`status <id> <status>`	Update non-verified task status.
`verify <id>`	Run checks and record verification evidence.

Command	Purpose
`related <id>`	Show dependencies and references.
`query`	Run deterministic graph queries.
`link suggest <id>`	Suggest code/test links.
`link update <id>`	Apply code/test/dependency/ref/gap link updates.
`link eval`	Run link-scorer fixtures.

Command	Purpose
`changed --since <ref>`	List committed, staged, unstaged, untracked, renamed, and deleted paths.
`drift --since <ref>`	Detect facts and warnings about task-memory drift.
`ci --since <ref>`	Produce local CI/PR packet.
`pr comment --since <ref>`	Generate review-only PR assistant output.
`repair --since <ref>`	Preview safe task-memory repairs.
`repair --apply`	Apply safe repairs.

Command	Purpose
`audit`	Validate task metadata and linked files.
`doctor`	Run diagnostics and repair suggestions.
`schema list`	List stable JSON schemas.
`schema show <name>`	Show one JSON schema.
`schema coverage`	Check schema catalog, artifacts, and SDK coverage.
`last`	Show the most recent mutation.
`history`	Show mutation history.
`handoff`	Summarize state for the next agent or human.

Most agent-facing commands support --json. Use schemas before writing scripts:

npx jumpspace schema list --json
npx jumpspace schema show work --json