locus config

View and manage Locus project settings stored in .locus/config.json.

View and update local project settings. Configuration is stored in .locus/config.json and is generated during locus init.

Usage

locus config <subcommand> [args]

If no subcommand is provided, show is used by default.

---

Subcommands

show

Display the current configuration organized by section.

locus config show
locus config           # Same as 'locus config show'

Sections displayed:

Section	Keys
GitHub	owner, repo, defaultBranch
AI	provider, model
Agent	maxParallel, autoLabel, autoPR, baseBranch, rebaseBeforeTask
Sprint	stopOnFailure
Logging	level, maxFiles, maxTotalSizeMB
Sandbox	enabled, extraWorkspaces, readOnlyPaths

set

Update a specific configuration value using dot-notation paths.

locus config set <path> <value>

Examples:

locus config set ai.model claude-sonnet-4-6
locus config set ai.provider codex
locus config set agent.maxParallel 5
locus config set agent.baseBranch develop
locus config set agent.rebaseBeforeTask false
locus config set sprint.stopOnFailure false
locus config set logging.level debug
locus config set sandbox.enabled false

get

Read a specific configuration value. Output goes to stdout for scripting.

locus config get <path>

Examples:

locus config get ai.model
locus config get sprint.stopOnFailure
locus config get agent.maxParallel

For nested objects, the output is formatted as JSON.

---

Configuration Keys

Path	Type	Default	Description
github.owner	string	(detected)	GitHub repository owner
github.repo	string	(detected)	GitHub repository name
github.defaultBranch	string	(detected)	Default branch (e.g., main)
ai.provider	string	"claude"	AI provider: claude or codex
ai.model	string	(provider default)	AI model identifier
agent.maxParallel	number	3	Maximum parallel issue execution
agent.autoLabel	boolean	true	Automatically manage labels during execution
agent.autoPR	boolean	true	Automatically create PRs after task completion
agent.baseBranch	string	(detected)	Branch that sprint and worktree branches are based on
agent.rebaseBeforeTask	boolean	true	Rebase sprint branch before each task
sprint.stopOnFailure	boolean	false	Stop sprint execution when a task fails (default: continue to next task)
logging.level	string	"normal"	Log level: silent, normal, verbose, debug
logging.maxFiles	number	10	Maximum number of log files to keep
logging.maxTotalSizeMB	number	50	Maximum total log size in MB
sandbox.enabled	boolean	true	Enable Docker sandbox isolation by default
sandbox.extraWorkspaces	string[]	[]	Additional workspace paths to sync into sandbox
sandbox.readOnlyPaths	string[]	[]	Paths mounted as read-only in sandbox

---

Examples

# Show all settings
locus config show

# Switch AI provider and model
locus config set ai.provider codex
locus config set ai.model gpt-5.3-codex

# Increase parallelism
locus config set agent.maxParallel 8

# Stop sprint on first failure (default is false — continue to next task)
locus config set sprint.stopOnFailure true

# Read a value for scripting
MODEL=$(locus config get ai.model)
echo "Using model: $MODEL"