agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Disable soul enforcement without deleting SOUL.md or soul files
$ npx -y skills add jmagly/aiwg --skill soul-disable --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/soul-disableContext preview
The summary Claude sees to decide when to auto-load this skill.
Disable soul enforcement without deleting SOUL.md or soul files
namespace: aiwg name: soul-disable platforms: [all] description: Disable soul enforcement without deleting SOUL.md or soul files commandHint: argumentHint: "[--provider claude|warp|copilot|cursor|factory|windsurf|opencode|codex] [--all]" allowedTools: Read, Write, Edit, Bash, Glob model: haiku category: soul-management modelRole: efficiency modelTier: economy
You are a Soul Management Specialist responsible for disabling SOUL.md enforcement.
Remove the soul directive from platform context file(s) and remove the enforcement rule so SOUL.md is NOT loaded at the next session start. The SOUL.md file itself is **preserved on disk** — no content is lost.
| Flag | Description | |------|-------------| | `--provider <name>` | Target specific provider: `claude`, `warp`, `copilot`, `cursor`, `factory`, `windsurf`, `opencode`, `codex` | | `--all` | Disable for all installed providers (default if no provider specified) |
If `--provider <name>` specified, operate on that provider only. If `--all` or no flag, detect installed providers by checking for their context files.
For each target provider:
1. Check if the context file exists 2. Check if the `@SOUL.md` directive is currently present
# Example for Claude Code grep -q "@SOUL.md" CLAUDE.md && echo "enabled" || echo "already disabled"
**For @-link style** (Claude, Warp, Windsurf, Copilot, Cursor, Factory, OpenCode):
Remove the `@SOUL.md` line from the context file. Preserve all surrounding content.
**For Codex** (inline injection):
Remove the block between soul injection markers:
<!-- BEGIN SOUL --> ...content... <!-- END SOUL -->
# Remove the enforcement rule rm -f .claude/rules/soul-enforcement.md
Check if any agent definitions have soul identity sections added by `soul-enable --agents`. If found, remove them:
# Check for soul-wired agents grep -rl "See @.*\.soul\.md" .claude/agents/ 2>/dev/null
Remove the `## Identity` section that references the soul file from each wired agent.
After removing directives and rule, confirm the soul file still exists:
ls SOUL.md .aiwg/SOUL.md 2>/dev/null && echo "soul file preserved" || echo "no soul file found"
Soul enforcement disabled Changes made: ~ CLAUDE.md (@SOUL.md directive removed) - .claude/rules/soul-enforcement.md (removed) SOUL.md preserved at ./SOUL.md Agent soul wiring preserved (agent definitions unchanged) To re-enable: /soul-enable To check status: /soul-status
If soul enforcement is already disabled:
Soul already disabled for Claude Code No @SOUL.md directive found in CLAUDE.md No enforcement rule found No changes made.
`soul-disable` NEVER:
It only removes the directive line that loads soul context and the enforcement rule.
| Condition | Action | |-----------|--------| | Context file missing | Report: nothing to disable | | SOUL.md missing | Report warning: soul disabled but SOUL.md not found on disk | | No write permission | Report permission error | | Enforcement rule missing | Skip removal, note already absent |
# Disable for all installed providers /soul-disable # Disable for Claude Code only /soul-disable --provider claude # Disable for Warp /soul-disable --provider warp
Reusable project context and specialist workflows for the AI tools you already use. Plan software, coordinate specialist reviews, prepare campaigns, investigate incidents, organize research, curate media, and maintain operational knowledge.
Repo: jmagly/aiwg
Crash-resilient external agent loop with state persistence and CI/CD integration
Detect requests for iterative autonomous agent loops and route to the appropriate loop executor
Automatically execute tests when code-generating agents modify source files, enforcing the execute-before-return pattern
Enable agent loops to learn from similar past tasks and share patterns across loops
Query and manage the executable feedback debug memory
Execute tests on generated code and iterate until passing