Skill

set-harness

Configures oh-my-harness settings for features like test enforcement, model routing, multi-agent workflows, thresholds, and scope guards via /set-harness.

developer-tools

npx claudepluginhub hoya324/oh-my-harness --plugin oh-my-harness

Tool Access

This skill uses the workspace's default tool permissions.

Preview

Configure oh-my-harness settings.

SKILL.md

Similar Skills

config

Configures Plan-Build-Run settings: workflow depth, model profiles, features, git branching/mode, and gates. Interactive or direct args like 'depth standard' or 'feature auto_continue on'.

5 tools

pbr

Claude Code Settings Deep Dive

Reference every Claude Code setting option from permissions, hooks, and sandbox to models, status line, and MCP servers. Use for config customization and troubleshooting.

claude-code-expert

harness-setup

830

Initializes Harness projects by generating CLAUDE.md, Plans.md, .claude/settings.json, configuring CI/Codex CLI/agents/memory, syncing skill mirrors, and localizing rules. Use for new projects, repairs, or maintenance.

6 tools

claude-code-harness

Stats

Stars5

Forks0

Last CommitMar 26, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Configure oh-my-harness settings.

Usage: /set-harness [setting.path value] Example: /set-harness testEnforcement.minCases 3 Example: /set-harness features.scopeGuard true Example: /set-harness multiAgent.useWorktree false

Steps

Read the current configuration from .claude/.omh/harness.config.json. If the file doesn't exist, suggest running /harness-setup first.
If no arguments provided, display all current values in a readable table format and exit.
If arguments are provided via $ARGUMENTS, parse them as <path> <value> and update the config.
Write the modified config back to .claude/.omh/harness.config.json and confirm the change.

Available Settings

features.conventionSetup (bool) — Auto-detect project conventions on session start
features.testEnforcement (bool) — Remind about tests after code changes
features.contextOptimization (bool) — Enable model routing for subagents
features.autoPlanMode (bool) — Suggest plan mode for multi-task requests
features.ambiguityDetection (bool) — Ask clarifying questions for vague requests
features.dangerousGuard (bool) — Warn before destructive commands
features.contextSnapshot (bool) — Save context before compaction
features.commitConvention (bool) — Remind commit message format
features.scopeGuard (bool) — Restrict file modifications to allowed paths
features.usageTracking (bool) — Track tool usage per session
features.autoGitignore (bool) — Auto-add .claude/.omh/ to .gitignore
features.skillScaffolding (bool) — Scaffold project-specific skills on init
testEnforcement.minCases (number) — Minimum test cases required per file
testEnforcement.promptOnMissing (bool) — Alert when tests are missing
modelRouting.quick/standard/complex (haiku/sonnet/opus) — Model for each tier
autoPlan.threshold (number) — Task count to trigger auto-plan
ambiguityDetection.threshold (number) — Ambiguity score to trigger clarification
commitConvention.style (auto/conventional/gitmoji) — Commit message format
scopeGuard.allowedPaths (string[]) — Directories allowed for modification
multiAgent.maxAgents (number) — Maximum parallel agents
multiAgent.useWorktree (bool) — Use isolated git worktrees per agent
multiAgent.tmuxSession (string) — Tmux session name for agents
features.nativeTeam (bool) — Enable native Claude Code team skills
nativeTeam.maxTeammates (number) — Maximum teammates per team
nativeTeam.defaultTeamName (string) — Default team name