From oh-my-god
Release the current Claude Code session from oh-my-god strict mode back to passive (default role). Lifts the PreToolUse tool allowlist but leaves Stop-hook gate judgment running.
How this skill is triggered — by the user, by Claude, or both
Slash command
/oh-my-god:omg-releaseThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Inverse of `omg-lock`. Puts the current session back into passive mode
Inverse of omg-lock. Puts the current session back into passive mode
by resetting session.role to default, so the PreToolUse scope hook
stops enforcing the role's tool allowlist and Claude can use any tool
again.
Important: release only lifts the tool boundary. The Stop-hook gate judgment (Day 4) still runs after every turn, so verification and retry-on-fail remain active.
The user invokes this when they want to relax the current session. Typical triggers:
Execute the release script via the Bash tool:
node "$CLAUDE_PLUGIN_ROOT"/scripts/omg-release.mjs
Read the stdout. It will report:
Summarize to the user:
If the script reports "already in default (passive) mode", just tell the user the session was already passive — no action needed.
User: /omg-release
Claude runs:
node "$CLAUDE_PLUGIN_ROOT"/scripts/omg-release.mjs
Script output:
[omg-release] session abc123 — role executor → default
[omg-release] tool boundary enforcement disabled.
[omg-release] Note: Stop-hook gate judgment still runs. Release only lifts the PreToolUse tool allowlist.
Claude to user:
Released session back to passive mode. The tool allowlist is no longer enforced. Stop-hook gate judgment still runs after every turn, so your work will still be verified. Say
/omg-lock <role>if you want strict mode back.
omg-lock, this defaults to the most recent session. Use
--session <id> to target a specific one if you're running
multiple Claude Code instances..claude/settings.json to
remove the hook registrations.npx claudepluginhub ckdwns9121/oh-my-god --plugin oh-my-godGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.