nemonix-cli 1.0.0

package/src/config/claude-code-config/skills/workflow-ci-fixer/references/vercel-cli.md

@@ -0,0 +1,192 @@
+# Vercel CLI Complete Reference
+
+<deployment_commands>
+## List Deployments
+```bash
+vercel ls                              # List all deployments
+vercel ls --prod                       # Production only
+vercel ls --meta key=value            # Filter by metadata
+vercel ls --output json               # JSON output
+vercel ls --scope team-slug           # Specific team
+```
+
+## Inspect Deployment
+```bash
+vercel inspect <deployment-url>        # Full deployment details
+vercel inspect <deployment-url> --json # JSON output
+
+# Output includes:
+# - Build command and output
+# - Environment variables used
+# - Functions and routes
+# - Error messages if failed
+```
+
+## Runtime Logs
+```bash
+vercel logs <deployment-url>           # Stream logs (5 min max)
+vercel logs <deployment-url> --json   # JSON format
+vercel logs <deployment-url> | jq 'select(.level == "warning")'  # Filter warnings
+```
+</deployment_commands>
+
+<build_commands>
+## Local Build
+```bash
+vercel build                           # Build with preview env
+vercel build --prod                    # Build with production env
+vercel build --debug                   # Verbose output
+vercel build --yes                     # Skip prompts, auto-pull env
+vercel build --target staging         # Custom environment
+```
+
+## Build Output
+Build artifacts are placed in `.vercel/output/` following Build Output API v3:
+```
+.vercel/output/
+├── config.json
+├── static/           # Static files
+└── functions/        # Serverless functions
+```
+</build_commands>
+
+<deploy_commands>
+## Deploy Options
+```bash
+vercel deploy                          # Deploy to preview
+vercel deploy --prod                   # Deploy to production
+vercel deploy --prebuilt              # Deploy pre-built output
+vercel deploy --force                  # Skip build cache
+vercel deploy --with-cache            # Keep cache with --force
+vercel deploy --logs                   # Show build logs
+vercel deploy --no-wait               # Don't wait for completion
+vercel deploy --archive tgz           # Compress for large projects
+vercel deploy -e KEY=value            # Runtime env var
+vercel deploy --build-env KEY=value   # Build-time env var
+vercel deploy --regions iad1          # Specify function region
+vercel deploy --target staging        # Custom environment
+vercel deploy --skip-domain           # Don't assign domains (with --prod)
+vercel deploy --meta key=value        # Add metadata
+```
+
+## Deploy from CI
+```bash
+# In CI, use token
+VERCEL_TOKEN=xxx vercel deploy --token=$VERCEL_TOKEN
+
+# Or with project linking
+vercel link --yes
+vercel deploy --yes
+```
+</deploy_commands>
+
+<environment_commands>
+## Environment Variables
+```bash
+vercel env ls                          # List all env vars
+vercel env ls production              # Production only
+vercel env add                        # Add interactively
+vercel env add SECRET production      # Add to specific env
+vercel env rm SECRET                  # Remove
+vercel env pull                       # Pull to .env.local
+vercel env pull .env                  # Pull to specific file
+```
+
+## Pull Project Settings
+```bash
+vercel pull                            # Pull env vars + settings
+vercel pull --yes                      # Skip confirmation
+vercel pull --environment production  # Specific environment
+```
+</environment_commands>
+
+<project_commands>
+## Project Management
+```bash
+vercel project ls                      # List projects
+vercel project add                     # Create project
+vercel project rm <name>              # Remove project
+vercel link                           # Link current directory
+vercel link --yes                     # Auto-confirm
+vercel unlink                         # Unlink project
+```
+
+## Domain Management
+```bash
+vercel domains ls                      # List domains
+vercel domains add <domain>           # Add domain
+vercel domains rm <domain>            # Remove domain
+vercel domains inspect <domain>       # Domain details
+vercel alias <deployment> <domain>    # Alias deployment
+```
+</project_commands>
+
+<debugging_commands>
+## Debug Build Issues
+```bash
+# 1. Pull latest settings
+vercel pull --yes
+
+# 2. Build locally with debug
+vercel build --debug 2>&1 | tee build.log
+
+# 3. Check for specific errors
+grep -i "error\|failed\|warning" build.log
+
+# 4. Check memory usage during build
+VERCEL_DEBUG_MEMORY_REPORT=1 vercel build
+```
+
+## Debug Runtime Issues
+```bash
+# Stream logs and filter errors
+vercel logs <url> --json | jq 'select(.level == "error")'
+
+# Get deployment events
+vercel inspect <url> --json | jq '.events'
+```
+
+## Force Fresh Build
+```bash
+# Skip all caches
+vercel deploy --force
+
+# Or set env var
+VERCEL_FORCE_NO_BUILD_CACHE=1 vercel deploy
+```
+</debugging_commands>
+
+<global_options>
+## Global Options (apply to all commands)
+```bash
+--cwd <dir>           # Set working directory
+--debug               # Debug output
+--global-config <dir> # Global config directory
+--help                # Show help
+--local-config <file> # Local config file
+--no-color            # Disable colors
+--scope <team>        # Team scope
+--token <token>       # Auth token
+```
+</global_options>
+
+<ci_integration>
+## CI/CD Integration
+```bash
+# GitHub Actions example
+- name: Deploy to Vercel
+  run: |
+    npm i -g vercel
+    vercel pull --yes --environment=production --token=${{ secrets.VERCEL_TOKEN }}
+    vercel build --prod --token=${{ secrets.VERCEL_TOKEN }}
+    vercel deploy --prebuilt --prod --token=${{ secrets.VERCEL_TOKEN }}
+
+# Or use official action
+- uses: amondnet/vercel-action@v25
+  with:
+    vercel-token: ${{ secrets.VERCEL_TOKEN }}
+    vercel-org-id: ${{ secrets.VERCEL_ORG_ID }}
+    vercel-project-id: ${{ secrets.VERCEL_PROJECT_ID }}
+    vercel-args: '--prod'
+```
+</ci_integration>

package/src/config/claude-code-config/skills/workflow-ci-fixer/steps/step-00-init.md

@@ -0,0 +1,157 @@
+---
+name: step-00-init
+description: Initialize ci-fixer workflow - parse flags, detect branch, setup state
+next_step: steps/step-01-watch-ci.md
+---
+
+# Step 0: Initialization
+
+## MANDATORY EXECUTION RULES (READ FIRST):
+
+- 🛑 NEVER skip flag parsing
+- ✅ ALWAYS detect the current git branch
+- 📋 Parse ALL flags before any other action
+- 💬 FOCUS on initialization only - don't start watching CI yet
+- 🚫 FORBIDDEN to proceed without proper state setup
+
+## EXECUTION PROTOCOLS:
+
+- 🎯 Parse flags first, then setup state
+- 💾 Initialize all state variables
+- 📖 Complete this step fully before loading next
+- 🚫 FORBIDDEN to load step-01 until init complete
+
+## YOUR TASK:
+
+Initialize the workflow by parsing flags, detecting the current branch, and setting up the execution environment.
+
+---
+
+## DEFAULTS CONFIGURATION:
+
+```yaml
+auto_mode: false       # -a, --auto: Skip confirmations
+max_attempts: 5        # -m N, --max-attempts=N: Max fix attempts
+current_attempt: 0     # Starts at 0
+fixes_applied: []      # Empty list
+local_verified: false  # Not verified yet
+```
+
+---
+
+## INITIALIZATION SEQUENCE:
+
+### 1. Parse Flags and Input
+
+**Step 1: Load defaults from config above**
+
+**Step 2: Parse user input and override defaults:**
+```
+Enable flags:
+  -a, --auto → {auto_mode} = true
+
+Set values:
+  -m N, --max-attempts=N → {max_attempts} = N
+```
+
+### 2. Detect Current Branch
+
+Run this command to get the current branch:
+```bash
+git branch --show-current
+```
+
+Store result in `{branch}`.
+
+**If no branch detected (detached HEAD):**
+→ Use `git rev-parse --short HEAD` as identifier
+→ Warn user: "Running in detached HEAD mode"
+
+### 3. Verify Git Status
+
+Run:
+```bash
+git status --porcelain
+```
+
+**If there are uncommitted changes:**
+→ Warn user: "You have uncommitted changes. They may interfere with CI fixes."
+
+**If `{auto_mode}` = false:**
+Use AskUserQuestion:
+```yaml
+questions:
+  - header: "Changes"
+    question: "You have uncommitted changes. How should we proceed?"
+    options:
+      - label: "Continue anyway (Recommended)"
+        description: "Proceed with uncommitted changes"
+      - label: "Stash changes"
+        description: "Stash changes before proceeding (git stash)"
+      - label: "Cancel"
+        description: "Stop and handle changes manually"
+    multiSelect: false
+```
+
+### 4. Initialize State
+
+Set all state variables:
+```yaml
+auto_mode: {parsed value}
+max_attempts: {parsed value}
+current_attempt: 0
+branch: {detected branch}
+run_id: null
+error_source: null
+error_logs: null
+fixes_applied: []
+local_verified: false
+```
+
+### 5. Confirm Start
+
+**If `{auto_mode}` = true:**
+→ Proceed directly to step-01
+
+**If `{auto_mode}` = false:**
+Use AskUserQuestion:
+```yaml
+questions:
+  - header: "Start"
+    question: "CI Fixer ready. Watch CI on branch '{branch}'?"
+    options:
+      - label: "Start watching (Recommended)"
+        description: "Begin monitoring CI pipeline"
+      - label: "Change settings"
+        description: "Modify configuration before starting"
+      - label: "Cancel"
+        description: "Don't start the workflow"
+    multiSelect: false
+```
+
+---
+
+## SUCCESS METRICS:
+
+✅ All flags correctly parsed
+✅ Current branch detected
+✅ Git status checked
+✅ State variables initialized
+✅ User confirmed start (if not auto_mode)
+
+## FAILURE MODES:
+
+❌ Not detecting the current branch
+❌ Proceeding with git errors
+❌ Missing state variables
+❌ Not warning about uncommitted changes
+
+---
+
+## NEXT STEP:
+
+After initialization, load `./step-01-watch-ci.md`
+
+<critical>
+Remember: Init is ONLY about setup - don't start watching CI here!
+</critical>

package/src/config/claude-code-config/skills/workflow-ci-fixer/steps/step-01-watch-ci.md

@@ -0,0 +1,192 @@
+---
+name: step-01-watch-ci
+description: Find CI run and monitor status - no upfront wait
+prev_step: steps/step-00-init.md
+next_step_success: steps/step-05-cleanup.md
+next_step_failure: steps/step-02-analyze-errors.md
+---
+
+# Step 1: Watch CI
+
+## MANDATORY EXECUTION RULES (READ FIRST):
+
+- 🛑 NEVER sleep upfront - check for CI immediately
+- ✅ ALWAYS check for CI run first, only wait if not found
+- 📋 YOU ARE A CI MONITOR, not a fixer (that comes later)
+- 💬 FOCUS on detecting CI status only
+- 🚫 FORBIDDEN to start fixing before CI has run
+
+## EXECUTION PROTOCOLS:
+
+- 🎯 Check immediately → Wait only if needed → Watch → Report
+- 💾 Store run_id when found
+- 📖 Complete monitoring before loading next step
+- 🚫 FORBIDDEN to load step-02 until CI status is known
+
+## CONTEXT BOUNDARIES:
+
+From step-00-init:
+
+| Variable | Description |
+|----------|-------------|
+| `{auto_mode}` | Skip confirmations |
+| `{max_attempts}` | Max fix attempts |
+| `{current_attempt}` | Current attempt number |
+| `{branch}` | Current git branch |
+| `{last_commit_sha}` | SHA of the commit we're watching |
+| `{artifacts_dir}` | Path to artifacts (may exist from previous attempts) |
+
+## YOUR TASK:
+
+Find the CI run for the current branch, wait only if not triggered yet (max 1 minute), then monitor until completion.
+
+---
+
+## EXECUTION SEQUENCE:
+
+### 1. Get Current Commit SHA
+
+First, capture the commit we're watching:
+```bash
+git rev-parse HEAD
+```
+Store as `{last_commit_sha}` (short: first 7 chars).
+
+### 2. Check for CI Run (Immediate - No Sleep)
+
+**Check immediately for a workflow run:**
+```bash
+gh run list --branch {branch} --limit 5 --json databaseId,status,conclusion,name,headSha
+```
+
+**Find run matching our commit:**
+- Look for a run where `headSha` starts with `{last_commit_sha}`
+- If found → Extract `databaseId` as `{run_id}`, go to step 3
+- If NOT found → Go to wait loop (step 2b)
+
+### 2b. Wait Loop (Only If No CI Found)
+
+**If no CI run found for our commit:**
+
+```
+wait_seconds = 0
+max_wait = 60  # 1 minute max
+
+WHILE wait_seconds < max_wait:
+    Display: "No CI run found yet. Waiting... ({wait_seconds}s / 60s)"
+
+    sleep 10
+    wait_seconds += 10
+
+    # Check again
+    gh run list --branch {branch} --limit 5 --json databaseId,status,conclusion,name,headSha
+
+    IF run found for {last_commit_sha}:
+        → Extract run_id, BREAK loop
+        → Go to step 3
+
+IF wait_seconds >= 60 AND still no run:
+    → Display: "⚠️ No CI configured for this repository"
+    → Display: "No GitHub Actions workflow triggered after 1 minute"
+    → Load step-05-cleanup.md (cleanup and exit)
+```
+
+### 3. Monitor Run Progress
+
+**If status = "in_progress" or "queued":**
+```bash
+gh run watch {run_id}
+```
+
+This command blocks until the run completes.
+
+**Display:**
+- "Watching run #{run_id} for commit {last_commit_sha}..."
+- Show job statuses as they complete
+
+### 4. Determine Outcome
+
+After run completes, check final status:
+```bash
+gh run view {run_id} --json conclusion,jobs
+```
+
+**If conclusion = "success":**
+→ Display: "✅ CI passed! All checks green."
+→ Load `step-05-cleanup.md` (cleanup and summary)
+
+**If conclusion = "failure":**
+→ Display: "❌ CI failed. Analyzing errors..."
+→ Increment `{current_attempt}`
+→ Proceed to step-02
+
+**If conclusion = "cancelled":**
+→ Display: "⚠️ CI was cancelled"
+→ Ask user what to do
+
+### 5. Check Attempt Limit
+
+**If `{current_attempt}` >= `{max_attempts}`:**
+→ Display: "Maximum attempts ({max_attempts}) reached."
+
+**If `{auto_mode}` = false:**
+Use AskUserQuestion:
+```yaml
+questions:
+  - header: "Max Attempts"
+    question: "Reached {max_attempts} attempts. What should we do?"
+    options:
+      - label: "Continue trying"
+        description: "Keep attempting to fix (reset counter)"
+      - label: "Stop here"
+        description: "Stop and show summary"
+    multiSelect: false
+```
+
+**If user chooses "Stop here" OR `{auto_mode}` = true:**
+→ Load `step-05-cleanup.md` (cleanup and summary)
+
+---
+
+## SUCCESS METRICS:
+
+✅ Checked for CI immediately (no upfront sleep)
+✅ Run ID captured for correct commit
+✅ Run monitored to completion
+✅ Status correctly determined (success/failure)
+✅ Attempt counter updated
+✅ Detected "no CI" after 1 minute if applicable
+✅ Routes to step-05 on success or stop
+
+## FAILURE MODES:
+
+❌ Sleeping before checking for CI
+❌ Waiting longer than 1 minute for non-existent CI
+❌ Missing run ID
+❌ Not matching run to correct commit SHA
+❌ Not incrementing attempt counter
+❌ Not routing to cleanup on success
+
+## WATCH PROTOCOLS:
+
+- Check immediately, don't assume CI needs time
+- Only wait if no run found (CI might be slow to trigger)
+- Max wait: 1 minute, then conclude no CI configured
+- Always match run to our specific commit SHA
+- Use `gh run watch` for real-time monitoring
+
+---
+
+## NEXT STEP:
+
+**If CI passed:** Load `./step-05-cleanup.md`
+
+**If no CI configured:** Load `./step-05-cleanup.md`
+
+**If max attempts reached and user stops:** Load `./step-05-cleanup.md`
+
+**If CI failed and attempts remaining:** Load `./step-02-analyze-errors.md`
+
+<critical>
+Remember: Check first, wait only if needed. Route to cleanup (step-05) on success or stop!
+</critical>