Interactive Onboarding and Proactive Update Notifications

[jdugarte]

🏷️ Type of Change

• 🚀 New Feature
• 🐛 Bug Fix
• 🧹 Refactor / Cleanup
• ⚡ Performance Optimization
• 📚 Documentation Update
• 🛠️ CI/CD / Build System

📝 Summary

This PR introduces a robust onboarding experience for new users and a proactive update notification system for existing users. It establishes an "Anti-Vibe-Coding" safeguard that guides users who write unstructured prompts towards using strict architectural workflows. It also leverages the AI's natural conversation during standard workflow GPS checks (explore-task, start-task, status-check) to passively alert the user if a new version of agentic:guild is available upstream without relying on background cron jobs.

✨ Key Changes

• "Who are you?" Onboarding Skill: Created skills/hello/SKILL.md to introduce the agent, assess the project's health by checking for core architectural documents, and provide a menu of copy-paste triggers.
• Terminal Call-to-Action: Updated sync.sh to print a highly visible call-to-action recommending the user trigger the new hello skill upon installation completion.
• Anti-Vibe Coding Safeguard: Added an <onboarding> directive to templates/core/AGENTIC_GUILD_RULES.md that intercepts raw, unstructured prompts and routes the user into either explore-task or start-task.
• Proactive Update Notifications: Added an <update_check> hook inside the pre-flight sequence for skills/explore-task/SKILL.md, skills/start-task/SKILL.md, and skills/status-check/SKILL.md. This silently fetches the upstream package.json and compares it to a localized copy of the package (.agenticguild/source_package.json), dropping an update banner into the chat if the upstream is higher or the local file is missing.
• Sync Registry Expansion: Updated playbooks/SYNC_REGISTRY.md to track the package.json and map the new hello skill for future synchronized updates.
• Bugbot Fixes: Standardized the hello skill XML schema to perfectly match the <agentic_guild_skill> template across other skills and patched a logic flaw in the missing file handling for version updates.

📸 Screenshots / Videos (if applicable)

N/A

🧪 Verification Plan

Automated Tests

• N/A

Manual Verification

1. [User opens a fresh project and runs sync.sh: verifies they see the output urging them to type "Who are you?"]
2. [User types "Who are you?": verifies the hello skill executes and runs the structural document checks.]
3. [User types a raw prompt like "build a login page": verifies AGENTIC_GUILD_RULES.md correctly steps in and routes them to explore-task or start-task.]
4. [User runs status-check, explore-task, or start-task with a missing or outdated local version of package.json: verifies the update notification appears natively in the response.]

🛡️ Impact Assessment

• Security: No impact
• Performance: No impact
• Accessibility: No impact

🧠 Anti-Drift Checklist

To prevent "drift" between the codebase and our AI context, please attest that you have updated the authoritative documentation:

• Docs: I have run sync-docs and updated any docs that require changes (SPEC, schema, DATA_FLOW_MAP, ADRs, etc.).
• Roadmap: If this PR completes a roadmap item, docs/ROADMAP.md has been updated (finish-branch does this automatically).
• Architecture (ADR): I have created/updated ADRs in docs/core/ADRs/ if new paradigms were introduced.
• AI Rules: I have updated .cursorrules or approved a harvest-rules run if developer protocols changed.
• Boundaries: I have updated docs/core/SYSTEM_ARCHITECTURE.md if the tech stack, libraries, or core patterns shifted.

🔍 Reviewer Focus Areas

Verify that all XML updates in the <update_check> hooks explicitly instruct the AI not to block the underlying workflow task execution.

🔜 Follow-ups / PENDING

None

[cursor]

PR Summary

Low Risk
Low risk: changes are limited to agentic:guild workflow docs/skills and installer messaging, with no application/runtime code paths affected. Main risk is behavioral (new intent routing and optional version-check notices) potentially altering how the assistant responds to prompts.

Overview
Adds a new hello skill that introduces agentic:guild, checks for required core docs, and provides copy/paste next-step commands; it’s now synced via SYNC_REGISTRY.md and routed from .cursorrules when users ask “who are you/what is agentic guild/how do I start”.

Adds an upstream update check to start-task, explore-task, and status-check, comparing package.json version from GitHub to a locally-synced .agenticguild/source_package.json and displaying a “new version available” notice when applicable.

Updates sync.sh to print a post-install onboarding hint, and extends onboarding rules in AGENTIC_GUILD_RULES.md to steer unstructured prompts into start-task/explore-task (anti-vibe-coding safeguard).

^{Written by Cursor Bugbot for commit 5a70406. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}