feat: add aidevops skill CLI command with telemetry disabled

.agent/AGENTS.md

@@ -511,7 +511,7 @@ Subagents provide specialized capabilities. Read them when tasks require domain
 | `workflows/` | Development processes - branching, releases, PR reviews, quality gates | git-workflow, plans, release, version-bump, pr, review-issue-pr, preflight, postflight, ralph-loop, session-review |
 | `templates/` | Document templates - PRDs, task lists, planning documents | prd-template, tasks-template, plans-template, todo-template |
 | `workflows/branch/` | Branch conventions - naming, purpose, merge strategies per branch type | feature, bugfix, hotfix, refactor, chore, experiment, release |
-| `scripts/commands/` | Slash commands - save-todo, remember, recall, code-simplifier, humanise and other interactive commands | save-todo, remember, recall, code-simplifier, humanise |
+| `scripts/commands/` | Slash commands - save-todo, remember, recall, code-simplifier, humanise, add-skill and other interactive commands | save-todo, remember, recall, code-simplifier, humanise, add-skill |
 
 <!-- AI-CONTEXT-END -->
 
@@ -554,6 +554,7 @@ aidevops features                # List available features
 | `aidevops upgrade` | Alias for update |
 | `aidevops repos` | List registered projects |
 | `aidevops repos add` | Register current project |
+| `aidevops skill <cmd>` | Manage imported skills (add/list/check/update/remove/generate) |
 | `aidevops detect` | Find unregistered aidevops projects |
 | `aidevops update-tools` | Check for outdated tools |
 | `aidevops uninstall` | Remove aidevops |
@@ -600,6 +601,7 @@ For AI-assisted setup guidance, see `aidevops/setup.md`.
 | Programmatic video | `tools/video/remotion.md` (React video creation, animations) |
 | AI image/video generation | `tools/video/higgsfield.md` (100+ generative models via unified API) |
 | Mobile emulators | `tools/mobile/minisim.md` (iOS simulator, Android emulator launcher) |
+| Importing skills | `scripts/commands/add-skill.md` (import, naming, update tracking) |
 
 ## Security
 
@@ -644,6 +646,47 @@ Never create files in `~/` root for files needed only with the current task.
 | `full-loop-helper.sh` | End-to-end development loop (task → PR → deploy) |
 | `session-review-helper.sh` | Gather session context for completeness review |
 | `humanise-update-helper.sh` | Check for upstream updates to humanise subagent |
+| `add-skill-helper.sh` | Import external skills from GitHub repos |
+| `skill-update-helper.sh` | Check/update imported skills from upstream |
+| `generate-skills.sh` | Generate SKILL.md stubs for cross-tool discovery |
+
+## Imported Skills
+
+Import community [Agent Skills](https://agentskills.io) into aidevops with upstream tracking.
+
+**Naming convention**: Imported skills use a `-skill` suffix to distinguish from native subagents:
+
+| Type | Pattern | Example | Managed by |
+|------|---------|---------|------------|
+| Native subagent | `{name}.md` | `playwright.md` | aidevops team |
+| Imported skill | `{name}-skill.md` | `playwright-skill.md` | Upstream repo |
+
+**CLI commands:**
+
+```bash
+aidevops skill add <source>       # Import from GitHub (→ *-skill.md)
+aidevops skill list                # List imported skills
+aidevops skill check               # Check for upstream updates
+aidevops skill update [name]       # Pull upstream changes
+aidevops skill remove <name>       # Remove imported skill
+aidevops skill generate            # Generate SKILL.md stubs for cross-tool discovery
+aidevops skill clean               # Remove generated SKILL.md stubs
+```
+
+**How it works:**
+1. Clones the source repo, detects format (SKILL.md, AGENTS.md, .cursorrules)
+2. Converts to aidevops subagent format with `-skill.md` suffix
+3. Places in appropriate category folder (auto-detected from content)
+4. Registers in `.agent/configs/skill-sources.json` for update tracking
+5. Telemetry disabled - no data sent to skills.sh or third parties
+
+**When issues arise with a subagent:**
+- `*-skill.md` → Check upstream for updates: `aidevops skill check`
+- Native `.md` → Evolve locally, submit PR to aidevops
+
+**Cross-tool discovery**: `aidevops skill generate` creates lightweight SKILL.md stubs so other tools (Cursor, VS Code Copilot, OpenCode) can discover aidevops subagents without changing our file structure.
+
+**Related scripts:** `add-skill-helper.sh`, `skill-update-helper.sh`, `generate-skills.sh`
 
 ## Quality Workflow
 

.agent/scripts/add-skill-helper.sh

@@ -283,11 +283,14 @@ determine_target_path() {
         category="services/hosting"
     fi
 
-    echo "$category/$skill_name"
+    # Append -skill suffix to distinguish imported skills from native subagents
+    # This enables: glob *-skill.md for imports, update checks, conflict avoidance
+    echo "$category/${skill_name}-skill"
     return 0
 }
 
 # Check for conflicts with existing files
+# Returns conflict info with type: NATIVE (our subagent) or IMPORTED (previous skill)
 check_conflicts() {
     local target_path="$1"
     local agent_dir="$2"
@@ -299,11 +302,28 @@ check_conflicts() {
     local conflicts=()
 
     if [[ -f "$md_path" ]]; then
-        conflicts+=("$md_path")
+        if [[ "$md_path" == *-skill.md ]]; then
+            conflicts+=("IMPORTED: $md_path")
+        else
+            conflicts+=("NATIVE: $md_path")
+        fi
     fi
 
     if [[ -d "$dir_path" ]]; then
-        conflicts+=("$dir_path/")
+        if [[ "$dir_path" == *-skill ]]; then
+            conflicts+=("IMPORTED: $dir_path/")
+        else
+            conflicts+=("NATIVE: $dir_path/")
+        fi
+    fi
+
+    # Also check for native subagent without -skill suffix (same base name)
+    local base_name="${target_path%-skill}"
+    local native_md="${agent_dir}/${base_name}.md"
+    if [[ "$target_path" == *-skill && -f "$native_md" ]]; then
+        # Native subagent exists with same base name - not a conflict since
+        # -skill suffix differentiates, but inform the user
+        conflicts+=("INFO: Native subagent exists at $native_md (no conflict, -skill suffix differentiates)")
     fi
 
     if [[ ${#conflicts[@]} -gt 0 ]]; then
@@ -491,8 +511,8 @@ cmd_add() {
             local skill_name="${custom_name:-$(basename "${subpath:-$repo}")}"
             skill_name=$(to_kebab_case "$skill_name")
             # openskills installs to ~/.config/opencode/skills/<name>/SKILL.md
-            # Register with .md extension for consistency with other paths
-            register_skill "$skill_name" "https://github.com/$owner/$repo" ".agent/skills/${skill_name}.md" "skill-md" "" "openskills" "Installed via openskills CLI"
+            # Register with -skill suffix for consistency with direct imports
+            register_skill "$skill_name" "https://github.com/$owner/$repo" ".agent/skills/${skill_name}-skill.md" "skill-md" "" "openskills" "Installed via openskills CLI"
             return 0
         fi
         log_warning "openskills failed, falling back to direct fetch"
@@ -558,35 +578,62 @@ cmd_add() {
     local conflicts
     conflicts=$(check_conflicts "$target_path" ".agent") || true
     if [[ -n "$conflicts" ]]; then
-        if [[ "$force" != true ]]; then
-            log_warning "Conflicts detected:"
-            echo "$conflicts" | while read -r conflict; do
-                echo "  - $conflict"
+        # Filter out INFO lines (informational, not blocking)
+        local blocking_conflicts
+        blocking_conflicts=$(echo "$conflicts" | grep -v "^INFO:" || true)
+        local info_lines
+        info_lines=$(echo "$conflicts" | grep "^INFO:" || true)
+
+        # Show info lines (native subagent coexistence)
+        if [[ -n "$info_lines" ]]; then
+            echo "$info_lines" | while read -r info; do
+                log_info "${info#INFO: }"
             done
-            echo ""
+        fi
+
+        # Handle blocking conflicts
+        if [[ -n "$blocking_conflicts" && "$force" != true ]]; then
+            # Determine conflict type for better messaging
+            if echo "$blocking_conflicts" | grep -q "^NATIVE:"; then
+                log_warning "Conflicts with native aidevops subagent(s):"
+                echo "$blocking_conflicts" | while read -r conflict; do
+                    echo "  - ${conflict#NATIVE: }"
+                done
+                echo ""
+                echo "The -skill suffix should prevent this. If you see this,"
+                echo "the imported skill has the same name as a native subagent."
+                echo ""
+            elif echo "$blocking_conflicts" | grep -q "^IMPORTED:"; then
+                log_warning "Conflicts with previously imported skill(s):"
+                echo "$blocking_conflicts" | while read -r conflict; do
+                    echo "  - ${conflict#IMPORTED: }"
+                done
+                echo ""
+            else
+                log_warning "Conflicts detected:"
+                echo "$blocking_conflicts" | while read -r conflict; do
+                    echo "  - $conflict"
+                done
+                echo ""
+            fi
+
             echo "Options:"
-            echo "  1. Merge (add new content to existing)"
-            echo "  2. Replace (overwrite existing)"
-            echo "  3. Separate (use different name)"
-            echo "  4. Skip (cancel import)"
+            echo "  1. Replace (overwrite existing)"
+            echo "  2. Separate (use different name)"
+            echo "  3. Skip (cancel import)"
             echo ""
-            read -rp "Choose option [1-4]: " choice
+            read -rp "Choose option [1-3]: " choice
 
             case "$choice" in
                 1)
-                    log_info "Merging with existing..."
-                    # TODO: Implement merge logic
-                    log_warning "Merge not yet implemented, using replace"
-                    ;;
-                2)
                     log_info "Replacing existing..."
                     ;;
-                3)
+                2)
                     read -rp "Enter new name: " new_name
                     skill_name=$(to_kebab_case "$new_name")
                     target_path=$(determine_target_path "$skill_name" "$description" "$source_dir")
                     ;;
-                4|*)
+                3|*)
                     log_info "Import cancelled"
                     return 0
                     ;;
@@ -738,7 +785,7 @@ cmd_check_updates() {
         # Get latest commit from GitHub API
         local api_url="https://api.github.com/repos/$owner/$repo/commits?per_page=1"
         local latest_commit
-        latest_commit=$(curl -s "$api_url" | jq -r '.[0].sha // empty' 2>/dev/null)
+        latest_commit=$(curl -s --connect-timeout 10 --max-time 30 "$api_url" | jq -r '.[0].sha // empty' 2>/dev/null)
 
         if [[ -z "$latest_commit" ]]; then
             log_warning "Could not fetch latest commit for $name"

.agent/scripts/commands/add-skill.md

@@ -11,14 +11,17 @@ URL/Repo: $ARGUMENTS
 ## Quick Reference
 
 ```bash
-# Import skill from GitHub
+# Import skill from GitHub (saved as *-skill.md)
 /add-skill dmmulroy/cloudflare-skill
+# → .agent/services/hosting/cloudflare-skill.md
 
 # Import specific skill from multi-skill repo
 /add-skill anthropics/skills/pdf
+# → .agent/tools/pdf-skill.md
 
 # Import with custom name
 /add-skill vercel-labs/agent-skills --name vercel-deploy
+# → .agent/tools/deployment/vercel-deploy-skill.md
 
 # List imported skills
 /add-skill list
@@ -27,6 +30,21 @@ URL/Repo: $ARGUMENTS
 /add-skill check-updates
 ```
 
+## Naming Convention
+
+Imported skills are saved with a `-skill` suffix to distinguish them from native aidevops subagents:
+
+| Type | Example | Managed by |
+|------|---------|------------|
+| Native subagent | `playwright.md` | aidevops team, evolves with framework |
+| Imported skill | `playwright-skill.md` | Upstream repo, checked for updates |
+
+This means:
+- No name clashes between imports and native subagents
+- `*-skill.md` glob finds all imports instantly
+- `aidevops skill check` knows which files to check for upstream updates
+- Issues with imported skills → check upstream; issues with native → evolve locally
+
 ## Workflow
 
 ### Step 1: Parse Input
@@ -111,7 +129,7 @@ Imported skills are tracked in `.agent/configs/skill-sources.json`:
       "name": "cloudflare",
       "upstream_url": "https://github.com/dmmulroy/cloudflare-skill",
       "upstream_commit": "abc123...",
-      "local_path": ".agent/services/hosting/cloudflare.md",
+      "local_path": ".agent/services/hosting/cloudflare-skill.md",
       "format_detected": "skill-md",
       "imported_at": "2026-01-21T00:00:00Z",
       "last_checked": "2026-01-21T00:00:00Z",

.agent/tools/browser/playwriter.md

@@ -59,12 +59,14 @@ Add to your MCP client configuration:
   "mcp": {
     "playwriter": {
       "type": "local",
-      "command": ["npx", "playwriter@latest"],
+      "command": ["/opt/homebrew/bin/npx", "-y", "playwriter@latest"],
       "enabled": true
     }
   }
 }
-```text
+```
+
+> **Note**: Use full path to `npx` (e.g., `/opt/homebrew/bin/npx` on macOS with Homebrew) for reliability. The `-y` flag auto-confirms package installation.
 
 **Claude Desktop** (`claude_desktop_config.json`):
 
@@ -73,11 +75,30 @@ Add to your MCP client configuration:
   "mcpServers": {
     "playwriter": {
       "command": "npx",
-      "args": ["playwriter@latest"]
+      "args": ["-y", "playwriter@latest"]
     }
   }
 }
-```text
+```
+
+**Enable per-agent** (OpenCode tools section):
+
+```json
+{
+  "tools": {
+    "playwriter_*": false
+  },
+  "agent": {
+    "Build+": {
+      "tools": {
+        "playwriter_*": true
+      }
+    }
+  }
+}
+```
+
+> **Tip**: Disable globally with `"playwriter_*": false` in `tools`, then enable per-agent to reduce context token usage.
 
 ## Usage
 

CHANGELOG.md

@@ -7,11 +7,18 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- add `aidevops skill` CLI command for managing agent skills with telemetry disabled (#154)
+- add `-skill` suffix convention for imported skills to distinguish from native subagents
+- add `generate`/`clean` subcommands for SKILL.md cross-tool discovery stubs
+
 ## [2.72.0] - 2026-01-22
 
 ### Added
 
 - add MiniSim iOS/Android emulator launcher support (#151)
+
 ## [2.71.0] - 2026-01-22
 
 ### Added