boost command reference

All 76 commands, generated from the CLI itself — run boost <command> --help for the same in your terminal.

Install & lifecycle

installpkg

Install a skill from a tap registry

boost install [--force] [--agent A] [--scope {user,project}] [--dry-run] NAME [NAME ...]
Arguments
NAMEskill name, optionally qualified as tap:skill
Options
--forcereinstall even if already installed or pinned
--agent Alink only into this agent (repeatable)
--scope {user,project}rules/workflows: install into your user config (default) or the current repo (project)
--dry-runshow what would happen without changing anything

uninstallpkg

Remove an installed skill, rule, workflow, or config

boost uninstall NAME [NAME ...]
Arguments
NAME

syncpkg

Reconcile installed skills & symlinks against the lock file

boost sync [--diff] [--prune] [--json]
Options
--diffshow the plan without applying it
--prunealso delete orphaned store dirs
--jsonmachine-readable output

updatepkg

Sync taps or update installed skills

boost update [--taps-only] [TAP]
Arguments
TAPrefresh only this tap
Options
--taps-onlyrefresh tap clones & catalogs without touching skills

reinstallpkg

Reinstall a skill or all skills (force)

boost reinstall [--all] [NAME ...]
Arguments
NAME
Options
--allreinstall every installed skill

bundlepkg

Export/install skill sets via a Boostfile

boost bundle {dump,install} [FILE]
Arguments
{dump,install}
FILEBoostfile (dump: default stdout; install: default ./Boostfile, '-' = stdin)

importpkg

Import skills from a GitHub URL or local path

boost import [--name N] [--all] [--agent A] URL_OR_PATH
Arguments
URL_OR_PATH
Options
--name Nskill to pick when several are found (or a rename)
--allimport every skill found
--agent Alink only into this agent (repeatable)

migratepkg

Migrate skills between agents or from Skills CLI

boost migrate [--from AGENT] [--to AGENT] [--from-skills-cli] [--path DIR]
Options
--from AGENTagent migrating away from
--to AGENTagent to link every installed skill into
--from-skills-cliimport skills installed by the Skills CLI
--path DIRSkills CLI directory (default ~/.skills)

pinpkg

Pin a skill to its current version

boost pin NAME
Arguments
NAME

unpinpkg

Allow a pinned skill to update again

boost unpin NAME
Arguments
NAME

snapshotpkg

Save & restore whole skill environments

boost snapshot [--json] {save,list,restore} [LABEL|ID]
Arguments
{save,list,restore}
LABEL|IDlabel for save, snapshot id for restore
Options
--jsonmachine-readable output (list)

exportpkg

Package skills as shareable zip/tar archives

boost export [--out OUT] [--zip] [NAME ...]
Arguments
NAMEskills to export (default: all installed)
Options
-o, --out OUToutput archive path
--zipbuild a .zip instead of .tar.gz

adaptpkg

Render a skill as another framework's agent source (CrewAI, Agents SDK)

boost adapt [--to FRAMEWORK] [--model M] [--out FILE] name

Render a skill as another agent framework's native source

Arguments
nameskill to adapt
Options
--to FRAMEWORKtarget framework: agents-sdk, crewai, langgraph
--model MLLM the exported agent runs on (default: boost's ai.model; pass 'none' for the framework's own default)
-o, --out FILEwrite to FILE instead of stdout

Find & search

reindexfind

Build/refresh the full-content search index

boost reindex [--force] [--dense] [--json]

Build or refresh the full-content search index

Options
--forcereindex every tap, ignoring cached commits
--densealso embed chunks into the opt-in dense vector store (needs the `rag` extra and an embeddings API key)
--jsonmachine-readable output

discoverfind

Browse & search the GitHub-wide skill discovery index

boost discover [--limit LIMIT] [--json] [query ...]
Arguments
queryfilter terms (repo/path substring)
Options
--limit LIMITmax rows (default 25)
--jsonmachine-readable output

recommendfind

Suggest skills based on your project's tech stack

boost recommend [--path PATH] [--limit LIMIT] [--json]
Options
--path PATHproject directory (default: cwd)
--limit LIMITmax suggestions (default 8)
--jsonmachine-readable output

browsefind

Interactive full-screen TUI with fuzzy search

boost browse

indexfind

Build the discovery registry via GitHub Code Search

boost index [--limit LIMIT]
Options
--limit LIMITmax skill files to index (default 300)

statsfind

Install statistics & trend for a single skill

boost stats [--json] name
Arguments
nameskill name
Options
--jsonmachine-readable output

countfind

Quick summary of installed / available / taps

boost count [--json]
Options
--jsonmachine-readable output

Inspect & explain

listinfo

List installed skills

boost list [--tag TAG] [--json]

List installed skills, rules and workflows

Options
--tag TAGonly show skills carrying this tag
--jsonmachine-readable output

infoinfo

Show detailed info about a skill

boost info [--json] name
Arguments
name
Options
--jsonmachine-readable output

catinfo

Print a skill or rule's contents

boost cat [--raw] name
Arguments
name
Options
--rawno styling even on a TTY

editinfo

Open a skill's SKILL.md in your editor

boost edit name
Arguments
name

previewinfo

Render a SKILL.md with rich formatting

boost preview name
Arguments
name

explaininfo

Explain what a skill does in plain English

boost explain name
Arguments
name

loginfo

Git log for a skill, or boost's activity log

boost log [--limit N] [--diagnostics] [--crashes] [name]
Arguments
nameskill to show upstream history for
Options
-n, --limit Nmax entries (default 20)
--diagnosticsshow boost's diagnostic log trail (not skill history)
--crasheslist recent crash reports

homeinfo

Open a skill's GitHub page in the browser

boost home [--print] name
Arguments
name
Options
--printprint the URL without opening a browser

depsinfo

Show dependency & conflict relationships

boost deps [--json] [name]
Arguments
nameskill to inspect (default: check all installed)
Options
--jsonmachine-readable output

taginfo

Custom labels for organizing skills

boost tag [--list] [--json] [name] [mods ...]
Arguments
nameinstalled skill
mods+tag to add, -tag to remove
Options
--listshow every tag and the skills carrying it
--jsonmachine-readable output

Taps & registries

taptap

Add a GitHub repo as a skill registry

boost tap [--defaults] [--catalog] [--type {skill,rule,workflow}] [--category CATEGORY] [--limit N] [--include-lists] [--dry-run] [--curated] [spec]
Arguments
specowner/repo, a git URL, or a local directory
Options
--defaultstap the recommended public registries
--catalogtap from the bundled curated registry catalog
--type {skill,rule,workflow}with --catalog: restrict to one item type
--category CATEGORYwith --catalog: restrict to one category
--limit Nwith --catalog: only the top N registries by est. size
--include-listswith --catalog: also tap awesome-list/index repos
--dry-runwith --catalog: print what would be tapped, tap nothing
--curatedmark the tap as curated (★ in listings)

untaptap

Remove a registry tap

boost untap [--force] name
Arguments
nametap name (owner/repo or short alias)
Options
-f, --forceskip the confirmation prompt

tapstap

List all configured registry taps

boost taps [--json]
Options
--jsonmachine-readable output

outdatedtap

Show skills with available updates

boost outdated [--json]
Options
--jsonmachine-readable output

AI-assisted

distillai

Merge multiple skills into one deduplicated skill

boost distill [--output NEWNAME] [--install] NAME [NAME ...]
Arguments
NAMEtwo or more skills to merge
Options
-o, --output NEWNAMEname for the merged skill (default: <first>-distilled)
--installinstall the merged skill instead of writing a file

simulateai

Preview how a skill would change Claude's behavior

boost simulate [--task TEXT] NAME
Arguments
NAME
Options
--task TEXTtask to simulate (default: a typical coding task)

inferai

Generate a SKILL.md from your codebase patterns

boost infer [--path DIR] [--name N] [--output FILE] [--install]
Options
--path DIRcodebase to analyze (default: current directory)
--name Nskill name (default: project-conventions)
-o, --output FILEwrite to a file
--installinstall the generated skill

absorbai

Turn recurring chat-history patterns into a skill

boost absorb [--history PATH] [--limit N] [--install]
Options
--history PATHhistory .jsonl file or directory of them
--limit Nmax patterns to absorb (default: 5)
--installinstall the generated skill

evolveai

Iteratively improve a skill from feedback

boost evolve [--feedback TEXT] [--apply] NAME
Arguments
NAME
Options
--feedback TEXTwhat should change, in plain English
--applywrite the revision to the installed skill

contextai

Branch-aware skill activation

boost context status|enable|disable|map|unmap|apply
Arguments
status|enable|disable|map|unmap|apply

focusai

Temporarily prioritize skills for a work session

boost focus [--clear] [--status] [--json] [SKILL ...]
Arguments
SKILLskills to focus on
Options
--clearend the session and relink everything
--statusshow current focus
--json

impactai

Measure a skill's influence on code quality

boost impact [--json] [NAME]
Arguments
NAMEone skill (default: all installed)
Options
--json

Health & integrity

doctorchk

Check installation health & report issues

boost doctor

lintchk

Validate SKILL.md frontmatter & quality

boost lint [--tap TAP] [--min N] [--json] [NAME ...]
Arguments
NAME
Options
--tap TAPlint every skill in a tap's clone
--min Nminimum passing score (default 40)
--json

auditchk

Check installed skills against a safety blocklist

boost audit [--json]
Options
--json

verifychk

Validate skill quality & lock-file integrity

boost verify [--json] [NAME ...]
Arguments
NAME
Options
--json

driftchk

Detect installed skills diverging from source

boost drift [--json] [NAME ...]
Arguments
NAME
Options
--json

testchk

Validate installed skills against quality checks

boost test [NAME ...]
Arguments
NAME

fingerprintchk

Deterministic hash of the skill environment

boost fingerprint [--verbose] [--json]
Options
--verboseshow the hashed components
--json

quarantinechk

Isolate a problematic skill without uninstalling

boost quarantine [--release NAME] [--list] [NAME]
Arguments
NAME
Options
--release NAMEre-link a quarantined skill
--listlist quarantined skills

decaychk

Flag skills irrelevant to your current stack

boost decay [--json]
Options
--json

healchk

Self-diagnose & repair the boost environment

boost heal [--dry-run]
Options
--dry-runshow repairs without applying them

conflictchk

Detect contradictory rules between skills

boost conflict [--json]
Options
--json

changelogchk

Show a skill's upstream change history

boost changelog [-n N] NAME
Arguments
NAME
Options
-n Nnumber of entries (default 20)

attestchk

Display/verify the install record for skills

boost attest [--verify] [--json] [NAME]
Arguments
NAME
Options
--verifycheck sha & journal record for each skill
--json

healthchk

Dashboard of skill-environment health

boost health

Config & setup

configcfg

Display or modify boost configuration

boost config [--json] [{list,get,set,unset}] [key] [value]
Arguments
{list,get,set,unset}what to do (default: list)
keydotted key, e.g. ai.enabled
valuenew value (JSON or string)
Options
--jsonmachine-readable output

cleancfg

Clear stale caches & broken symlinks

boost clean [--dry-run] [--deep]
Options
--dry-runshow what would be removed without touching anything
--deepalso remove snapshots older than 90 days

createcfg

Scaffold a new skill from a template

boost create [--description DESCRIPTION] [--dir DIR] [--install] name
Arguments
nameskill name (slugified)
Options
--description DESCRIPTIONone-line trigger description for the frontmatter
--dir DIRparent directory (default: current directory)
--installinstall the new skill immediately

policycfg

Manage & enforce skill governance policies

boost policy [--json] [{list,set,unset,check}] [key] [value]
Arguments
{list,set,unset,check}what to do (default: list)
keypolicy key, e.g. min_quality_score
valuenew value (JSON, comma list, or string)
Options
--jsonmachine-readable output

onboardcfg

Add skill-tracker telemetry to a repo & open a PR

boost onboard [--repo REPO] [--pr] [--dry-run]
Options
--repo REPOrepository directory (default: .)
--prcommit on a branch and open a PR with `gh`
--dry-runpreview the files without writing anything

completionscfg

Generate shell tab-completion scripts

boost completions [{bash,zsh,fish}]
Arguments
{bash,zsh,fish}target shell (default: from $SHELL)

schedulecfg

Manage automatic skill-sync scheduling

boost schedule [--interval {6h,12h,daily}] [--json] [{status,enable,disable}]
Arguments
{status,enable,disable}what to do (default: status)
Options
--interval {6h,12h,daily}how often to run `boost update` (default: 6h)
--jsonmachine-readable output (status only)

servecfg

Serve the skill catalog over HTTP (port 8787)

boost serve [--port PORT] [--host HOST]
Options
--port PORTport to listen on (default: config serve.port)
--host HOSTaddress to bind (default: 127.0.0.1)

mcpcfg

Register boost as an MCP server for Claude Code

boost mcp [--stdio] [{register,unregister}]
Arguments
{register,unregister}what to do (default: register)
Options
--stdiorun the MCP server on stdin/stdout (used by Claude Code)

hookscfg

Manage Claude Code hooks (scope-aware) in settings.json

boost hooks [--command COMMAND] [--name NAME] [--scope {global,project}] [--matcher MATCHER] [--timeout TIMEOUT] {add,remove,list} [event]
Arguments
{add,remove,list}add | remove | list
eventhook event, e.g. SessionStart (required for add)
Options
-c, --command COMMANDcommand the hook runs (add)
-n, --name NAMEstable name used to tag & find the hook
-s, --scope {global,project}settings scope (default: project; list shows both)
-m, --matcher MATCHERClaude matcher, e.g. 'startup|resume|clear'
--timeout TIMEOUThook timeout in seconds (default: 10)

bmadcfg

Install & manage the BMAD Method (scope-aware, toggleable)

boost bmad [--scope {project,global}] [--modules MODULES] [--yes] [--startup] {install,init,startup,orient,uninstall,disable,enable,doctor,status} [value]
Arguments
{install,init,startup,orient,uninstall,disable,enable,doctor,status}install | init | startup | orient | uninstall | disable | enable | doctor | status
valuefor startup: on | off | status
Options
-s, --scope {project,global}target scope (default: project)
--modules MODULESBMAD modules to install (default: bmm)
-y, --yesskip confirmation prompts
--startupenable the startup toggle right after install

self-updatecfg

Update boost itself to the latest version

boost self-update

Team & sharing

cohortteam

Controlled skill rollouts & team A/B testing

boost cohort [--skills SKILLS] [--percent PERCENT] [--json] [{list,create,delete,status,apply}] [name]
Arguments
{list,create,delete,status,apply}
namecohort name
Options
--skills SKILLScomma-separated skill names
--percent PERCENTrollout percentage (default 100)
--jsonmachine-readable output

profileteam

Named skill profiles for context switching

boost profile [--prune] [--json] [{list,save,use,show,diff,delete}] [name]
Arguments
{list,save,use,show,diff,delete}
nameprofile name
Options
--prunewith `use`: fully uninstall skills not in the profile
--jsonmachine-readable output

protocolteam

Manage the boost:// one-click-install handler

boost protocol [{status,register,unregister,open}] [url]
Arguments
{status,register,unregister,open}
urla boost:// URL (for `open`)

pulseteam

Team activity feed of skill-management events

boost pulse [-n N] [--all] [--action ACTION] [--json]
Options
-n Nevents to show (default 20)
--allshow the whole journal
--action ACTIONfilter by action (install, tap, ...)
--jsonmachine-readable output

replayteam

View version history & roll back skills

boost replay [--json] [{list,show,rollback}] [id]
Arguments
{list,show,rollback}
idhistory entry id (from `boost replay list`)
Options
--jsonmachine-readable output

whoteam

Discover who on the team has skill expertise

boost who [--json] [skill]
Arguments
skillfocus on one skill
Options
--jsonmachine-readable output