@robbiesrobotics/alice-agents 1.2.0 → 1.2.2

package/lib/workspace-scaffolder.mjs

@@ -6,6 +6,7 @@ import { fileURLToPath } from 'node:url';
 const __dirname = dirname(fileURLToPath(import.meta.url));
 const TEMPLATES_DIR = join(__dirname, '..', 'templates', 'workspaces');
 const OPENCLAW_DIR = join(homedir(), '.openclaw');
+const SKILLS_DIR = join(OPENCLAW_DIR, 'skills');
 
 // Product files — always overwritten on install/upgrade
 const PRODUCT_FILES = ['SOUL.md', 'AGENTS.md', 'IDENTITY.md', 'TOOLS.md'];
@@ -99,6 +100,22 @@ export function scaffoldWorkspace(agent, userInfo, agentCount) {
   return { workspaceDir, written, skipped };
 }
 
+export function scaffoldSkills() {
+  // Install the claude-code skill into ~/.openclaw/skills/
+  const claudeCodeSkillDir = join(SKILLS_DIR, 'claude-code');
+  mkdirSync(claudeCodeSkillDir, { recursive: true });
+
+  const skillTemplatePath = join(__dirname, '..', 'templates', 'skills', 'claude-code', 'SKILL.md');
+  const skillDestPath = join(claudeCodeSkillDir, 'SKILL.md');
+
+  if (existsSync(skillTemplatePath)) {
+    const content = readFileSync(skillTemplatePath, 'utf8');
+    writeFileSync(skillDestPath, content, 'utf8');
+    return true;
+  }
+  return false;
+}
+
 export function scaffoldAll(agents, userInfo) {
   const agentCount = agents.length;
   const results = [];
@@ -108,5 +125,8 @@ export function scaffoldAll(agents, userInfo) {
     results.push({ agent: agent.id, ...result });
   }
 
+  // Install skills
+  scaffoldSkills();
+
   return results;
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@robbiesrobotics/alice-agents",
-  "version": "1.2.0",
+  "version": "1.2.2",
   "description": "A.L.I.C.E. — 28 AI agents for OpenClaw. One conversation, one team.",
   "bin": {
     "alice-agents": "bin/alice-install.mjs"

package/templates/skills/claude-code/SKILL.md

@@ -0,0 +1,111 @@
+---
+name: claude-code
+description: 'Delegate implementation tasks to Claude Code. Use when: (1) building new features or apps, (2) porting/refactoring code across multiple files, (3) fixing bugs that require file exploration, (4) any coding task that needs read+write+exec across a codebase. NOT for: planning/research (do that yourself), simple single-line edits (use edit tool directly), reading files (use read tool). Always give Claude Code a working directory and a precise task. Use background:true for long tasks and poll for completion.'
+metadata:
+  {
+    "openclaw": { "emoji": "⚡", "requires": { "anyBins": ["claude"] } }
+  }
+---
+
+# Claude Code Skill
+
+Runs Claude Code (`claude` CLI) in `--print --permission-mode bypassPermissions` mode.
+No PTY needed. Fully autonomous — reads, writes, edits, runs shell commands.
+
+## When to use this skill
+
+Use Claude Code when the task requires:
+- Creating or modifying **multiple files**
+- Running **build/test/lint** commands to verify correctness
+- Exploring an unfamiliar codebase before writing
+- Any implementation work that would take you many tool calls
+
+## Basic pattern
+
+```
+exec:
+  workdir: /path/to/project
+  command: claude --permission-mode bypassPermissions --print 'YOUR TASK HERE'
+```
+
+## Background pattern (for long tasks)
+
+```
+exec:
+  workdir: /path/to/project
+  background: true
+  command: claude --permission-mode bypassPermissions --print 'YOUR TASK HERE.
+
+When done, run: openclaw system event --text "Done: brief summary" --mode now'
+```
+
+Then poll with `process(action=log, sessionId=...)` to check progress.
+
+## Writing good prompts for Claude Code
+
+1. **Start with context** — tell it what the project is and what exists
+2. **Be specific** — name exact files, functions, behaviors expected
+3. **State success criteria** — "run npm run build and it must pass"
+4. **End with notification** — always append the `openclaw system event` line for long tasks
+
+## Example: Feature build
+
+```
+exec:
+  workdir: /Users/aliceclaw/.openclaw/workspace-olivia/mission-control
+  background: true
+  command: >
+    claude --permission-mode bypassPermissions --print '
+    This is a Next.js 16 + TypeScript + shadcn/ui app.
+    
+    Task: Add an AgentWizard modal to src/app/agents/page.tsx.
+    - Read src/app/agents/page.tsx first
+    - Read src/components/ui/ to see available components
+    - Build a multi-step Dialog: name → domain → preview → confirm
+    - On confirm, POST /api/agents with the form data
+    - The page should refresh the agent list after creation
+    
+    Run npm run build at the end and fix any TypeScript errors.
+    
+    When done: openclaw system event --text "AgentWizard complete" --mode now
+    '
+```
+
+## Example: Bug fix
+
+```
+exec:
+  workdir: /path/to/project
+  command: claude --permission-mode bypassPermissions --print 'Fix the TypeScript error in src/lib/openclaw.ts line 42. Run npm run build to verify.'
+```
+
+## Example: Code review
+
+```
+exec:
+  workdir: /path/to/project
+  command: claude --permission-mode bypassPermissions --print 'Review src/app/agents/page.tsx for bugs, performance issues, and missing error handling. Output a numbered list of findings.'
+```
+
+## Rules for agents using this skill
+
+1. **Give a precise workdir** — Claude Code needs to know where it is
+2. **One task per invocation** — don't chain unrelated work in one prompt
+3. **Always verify** — ask Claude Code to run the build/tests before finishing
+4. **Background for >5min tasks** — use background:true and the notification line
+5. **Report results** — after Claude Code finishes, summarize what changed to Olivia
+6. **Never run in ~/.openclaw/ itself** — always in a project subdirectory
+
+## Agents authorized to use this skill
+
+All coding specialists: dylan, felix, isaac, caleb, darius, avery, devon, quinn, alex
+Orchestrator: olivia (for coordination tasks only)
+
+## Checking if claude is available
+
+```
+exec:
+  command: which claude && claude --version
+```
+
+If not found, tell Olivia — she'll handle escalation.