agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Show current AIWG customization status — mode, source path, what you've customized vs upstream
$ npx -y skills add jmagly/aiwg --skill customize-status --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/customize-statusContext preview
The summary Claude sees to decide when to auto-load this skill.
Show current AIWG customization status — mode, source path, what you've customized vs upstream
namespace: aiwg name: customize-status platforms: [all] description: Show current AIWG customization status — mode, source path, what you've customized vs upstream
You report the current AIWG customization status: whether the user is in customization mode, what path their AIWG runs from, how many customizations they've made, and whether upstream has new commits they haven't pulled yet.
| Pattern | Example | Action | |---------|---------|--------| | Mode check | "am I in customize mode?" | `aiwg version` check | | Changes summary | "what have I customized?" | git diff vs upstream | | Full status | "customization status" | Full report |
When triggered:
1. **Check current mode**:
aiwg version
Look for `[dev]` in output and the edgePath. If not in dev mode, report that and offer to run `customize-setup`.
2. **Load edgePath** from config (or parse from `aiwg version` output).
3. **Detect fork vs clone** — check if `upstream` remote exists:
git -C <edgePath> remote -v
4. **Summarize customizations** — files the user added or modified vs what's in the repo:
# If fork (upstream remote exists): git -C <edgePath> diff upstream/main --name-only -- agentic/code/ .claude/ # If local clone only: git -C <edgePath> status --short -- agentic/code/ .claude/
Group by type: rules, agents, skills, prompts.
5. **Check upstream staleness** (fork mode only):
git -C <edgePath> fetch upstream --dry-run 2>&1 | grep -c "^From" || echo "0" git -C <edgePath> rev-list HEAD..upstream/main --count
6. **Report** in plain language:
AIWG Customization Status
Mode: live (fork)
Source: ~/my-aiwg
Fork: github.com/user/aiwg
Upstream: github.com/jmagly/aiwg — 3 commits ahead
Your customizations (4 files):
rules/ my-conventions.md (3 days ago)
team-style.md (1 week ago)
agents/ domain-specialist.md (2 days ago)
skills/ my-shortcuts/SKILL.md (1 day ago)
→ Say "apply my changes" to redeploy
→ Say "sync my AIWG" to pull the 3 upstream commitsIf no customizations yet:
Your customizations: none yet Try: "add a rule that [something personal]" then "apply my changes"
**User**: "what have I customized?"
**Action**: `aiwg version` → `git remote -v` → `git diff upstream/main --name-only` → `git rev-list HEAD..upstream/main --count`
**Response**: Full status report as above.
**User**: "am I in customize mode?"
**Action**: `aiwg version`
**Response**: "Yes — AIWG is running from ~/my-aiwg [dev mode]. You have 2 customizations."
**User**: "customization status"
**Action**: `aiwg version` → not in dev mode
**Response**: "AIWG is running from the npm package (not in customization mode). Say 'set up AIWG customization mode' to get started."
None needed — report what you find.
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