agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Enable soul enforcement by wiring SOUL.md into platform context files and deploying the enforcement rule
$ npx -y skills add jmagly/aiwg --skill soul-enable --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/soul-enableContext preview
The summary Claude sees to decide when to auto-load this skill.
Enable soul enforcement by wiring SOUL.md into platform context files and deploying the enforcement rule
namespace: aiwg name: soul-enable platforms: [all] description: Enable soul enforcement by wiring SOUL.md into platform context files and deploying the enforcement rule commandHint: argumentHint: "[--provider claude|warp|copilot|cursor|factory|windsurf|opencode|codex] [--all] [--agents]" allowedTools: Read, Write, Edit, Bash, Glob model: haiku category: soul-management modelRole: efficiency modelTier: economy
You are a Soul Management Specialist responsible for enabling SOUL.md enforcement in platform context files.
Wire a project's SOUL.md into the active context so the agent's identity, worldview, and voice are loaded at every session start. This involves:
1. Adding an `@SOUL.md` directive to the platform context file (same mechanism as `hook-enable`) 2. Deploying the soul enforcement rule to `.claude/rules/` 3. Optionally wiring per-agent soul files
| Flag | Description | |------|-------------| | `--provider <name>` | Target specific provider: `claude`, `warp`, `copilot`, `cursor`, `factory`, `windsurf`, `opencode`, `codex` | | `--all` | Enable for all installed providers (default if no provider specified) | | `--agents` | Also wire per-agent `.soul.md` files into their agent definitions | | `--agent <name>` | Wire soul for a specific agent only |
SOUL.md is looked up in priority order:
1. `SOUL.md` (project root) 2. `.aiwg/SOUL.md`
Per-agent soul files follow the pattern: `<agent-name>.soul.md` alongside the agent definition.
| Provider | Context File | Directive Added | |----------|-------------|-----------------| | Claude Code | `CLAUDE.md` | `@SOUL.md` | | Warp Terminal | `WARP.md` | `@SOUL.md` | | Windsurf | `AGENTS.md` | `@SOUL.md` | | GitHub Copilot | `.github/copilot-instructions.md` | `@SOUL.md` | | Cursor | `.cursorrules` | `@SOUL.md` | | Factory AI | `AGENTS.md` | `@SOUL.md` | | OpenCode | `.opencode/context.md` | `@SOUL.md` | | Codex | `CODEX.md` | Inline injection (no @-link) |
# Check for SOUL.md in priority order ls SOUL.md .aiwg/SOUL.md 2>/dev/null
If no SOUL.md found:
Error: No SOUL.md found. Looked in: ./SOUL.md ./.aiwg/SOUL.md Create one with: /soul-create
Count tokens in the soul file to warn about context budget impact:
wc -w SOUL.md
If over ~3750 words (~5K tokens), warn:
Warning: SOUL.md is approximately {N} tokens — this exceeds the recommended 5K limit.
Large soul files strain context budget in multi-agent workflows.
Consider trimming or run /soul-enhance to identify vague sections.
Proceed anyway? [y/n]If `--provider <name>` specified, operate on that provider only. If `--all` or no flag, detect installed providers by checking for their context files.
ls CLAUDE.md WARP.md AGENTS.md .github/copilot-instructions.md .cursorrules CODEX.md 2>/dev/null ls .opencode/context.md 2>/dev/null
For each target provider:
1. Check if the context file exists 2. Check if `@SOUL.md` directive is already present
# Example for Claude Code grep -q "@SOUL.md" CLAUDE.md && echo "already enabled" || echo "disabled"
If directive is missing, add `@SOUL.md` to the context file.
**Placement**: After any `@AIWG.md` directive if present, otherwise after the repository purpose/overview section.
@AIWG.md @SOUL.md
**For Codex** (no @-link support): Insert SOUL.md content between markers:
<!-- BEGIN SOUL -->
{SOUL.md content}
<!-- END SOUL -->Write the soul enforcement rule to `.claude/rules/soul-enforcement.md`:
# Check if rule already exists ls .claude/rules/soul-enforcement.md 2>/dev/null
If missing, create it from the template at `agentic/code/addons/aiwg-utils/rules/soul-enforcement.md`.
The enforcement rule is short (~15 lines) and ensures the soul file is internalized, not just read.
If `--agents` flag is set:
1. Find all agent definitions in `.claude/agents/` 2. For each agent, check if a companion `.soul.md` file exists 3. If found, add an identity reference to the agent definition
# Find agent definitions with companion soul files
for agent in .claude/agents/*.md; do
name=$(basename "$agent" .md)
soul=".claude/agents/${name}.soul.md"
if [ -f "$soul" ]; then
echo "Found soul for: $name"
fi
done**Identity reference added to agent definition**:
## Identity
See .claude/agents/{name}.soul.md (deployed soul file) for this agent's character and voice.If `--agent <name>` is specified, only wire that single agent.
Soul enforcement enabled Changes made: + .claude/rules/soul-enforcement.md (created) ~ CLAUDE.md (@SOUL.md directive added) SOUL.md loaded: ./SOUL.md (~2,847 tokens) Enforcement rule: 15 lines (~50 tokens per session) To disable: /soul-disable To check status: /soul-status
If `--agents` was used:
Agent soul wiring: ~ .claude/agents/test-engineer.md (identity section added) ~ .claude/agents/security-auditor.md (identity section added) - .claude/agents/api-designer.md (no companion .soul.md found, skipped)
If soul enforcement is already enabled:
Soul already enabled for Claude Code @SOUL.md directive found in CLAUDE.md .claude/rules/soul-enforcement.md exists No changes made.
| Condition | Action | |-----------|--------| | No SOUL.md found | Fail with message: "No SOUL.md found. Run /soul-create first." | | SOUL.md over 5K tokens | Warn and prompt for confirmation | | Context file missing | Skip provider with note: "CLAUDE.md not found — run `aiwg use claude` first." | | No write permission | Report per
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