Menu
Fast git sync with project codes - add, commit, pull (no rebase), push
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
[WHAT] Standalone alignment auditor for north-star drift. [HOW] Scans codebase state, compares to canonical docs, classifies drift, proposes corrections. [WHEN] Use only for explicit deep-dive alignment audits. [WHY] Default alignment should run inside `work`; this remains for focused audit sessions. Triggers: "align lev", "check north star", "what's our direction", "audit alignment", "detect drift", "product pivot"
[WHAT] Builder workflow for developing `~/lev` and migrating proven work to core paths. [HOW] Assess existing code, run prior art checks, decide placement against `~/lev/docs`, apply patches, validate E2E. [WHEN] Use when formalizing POCs, placing modules, or migrating workshop output into production architecture. [WHY] Prevents drift and duplicate effort by enforcing file-based routing and docs-aligned placement. Triggers: "build feature", "migrate poc", "where should this go", "prior art check", "placement decision", "formalize"
Agentic UX design language and pattern library. Establishes CLI-as-Prompt paradigm where every CLI output is a prompt to the next agent. Includes Enriched Response Protocol, vernacular (Self-Injection, Teachable Touch, Bi-Directional), and implementation patterns for validation gates and trust calibration.
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
Agent-human interaction protocol for AI agents. Use when agents need human approvals, selections, research direction, or async feedback. Triggers on "human in the loop", "approval", "agent interaction", "get human input", or "wait for response".
Think and reason like a software architect. Analyze systems through quality attributes, trade-off analysis, and architectural drivers. Use for system design, tech selection, ADRs, architecture reviews, C4 descriptions, fitness functions, and PR review. Triggers: "architect this", "design system", "architecture review", "ADR", "trade-off analysis", "quality attributes", "fitness functions", "C4 diagram", "system design", "review PR", "review this", "/review".
| name | gitsync |
| description | Fast git sync with project codes - add, commit, pull (no rebase), push |
| trigger | /sync, gitsync, "sync repos", "git sync", "push all" |
Fast parallel git sync across repos with project code resolution.
# By project code
/sync naac lev clawd chezmoi # Sync named projects
/sync all # Sync all dirty repos
/sync all --dry-run # Preview what would sync
# By path
/sync ~/proj1 ~/proj2 # Sync specific paths
/sync . # Sync current directory
# Options
--dry-run # Preview without changes
The registry at ~/.claude/skills/gitsync/projects.json maps codes to paths:
| Code | Path |
|---|---|
naac | ~/k/apps/production/naac |
lev | ~/lev |
clawd | ~/clawd |
chezmoi | ~/.local/share/chezmoi |
dotfiles | ~/.dotfiles |
leviathan | ~/digital/leviathan |
clawdbot | ~/.clawdbot |
| Symbol | Meaning |
|---|---|
| b | bd flush (daemon running) |
| B | bd sync (no daemon) |
| C | Committed |
| P | Pulled |
| U | Pushed |
| . | Skipped (nothing to do) |
| D | Dirty (dry-run only) |
git add .git commit -m "sync MM-DD HH:MM" (if dirty)git pull --no-rebasegit push/sync all scans all registered projects and syncs those with uncommitted changes.
/sync all # Sync all dirty repos
/sync all --dry-run # List dirty repos without syncing
When the script exits with code 1 (conflicts detected), the agent MUST:
# For each conflicted repo reported by gitsync:
cd <conflicted-repo>
git status --porcelain | grep "^UU" # List conflicted files
git diff --name-only --diff-filter=U # Alternative
Use AskUserQuestion with these options:
header: "Merge Conflict"
question: "How should I resolve conflicts in <repo>?"
options:
- "Accept ours (keep local changes)"
- "Accept theirs (take remote changes)"
- "Smart merge (launch subagent)"
- "Skip (I'll handle manually)"
Accept ours:
git checkout --ours .
git add .
git commit -m "resolve: accept local changes"
git push
Accept theirs:
git checkout --theirs .
git add .
git commit -m "resolve: accept remote changes"
git push
Smart merge (subagent):
Launch Task agent with subagent_type="code-refactoring:code-reviewer"
Prompt: "Analyze and resolve merge conflicts in <repo>.
Show me the conflicts, explain what each side changed,
and propose a merged solution that preserves both intents."
After resolution, re-run gitsync <repo> to complete the push.
If conflicts detected:
cd <conflicted-repo>
git status # see conflicts
git checkout --ours <file> # take local
git checkout --theirs <file> # take remote
git add . && git commit # complete merge
Then re-run /sync to continue.
~/.claude/skills/gitsync/scripts/gitsync.sh [codes|paths...] [--dry-run]
ln -sf ~/.claude/skills/gitsync/scripts/gitsync.sh ~/.claude/bin/gitsync
Skill-specific technique rationale. Apply patterns from the skill body. Progressive disclosure: metadata first, body on trigger, references on demand.
Role Definition: Specialist for gitsync domain. Executes workflows, produces artifacts, routes to related skills when needed.
Input Contract: Context, optional config, artifacts from prior steps. Depends on skill.
Output Contract: Artifacts, status, next-step recommendations. Format per skill.
Edge Cases & Fallbacks: Missing context—ask or infer from workspace. Dependency missing—degrade gracefully; note in output. Ambiguous request—clarify before proceeding.