Replace bmad-init dependency with direct config loading

[bmadcode]

Summary

• Agent-builder and workflow-builder skills now load config directly from {project-root}/_bmad/bmb/config.yaml instead of invoking the bmad-init skill
• Makes both skills self-contained — no external skill dependency needed for initialization
• Explicitly lists resolved config variables (user_name, communication_language, bmad_builder_output_folder, bmad_builder_reports)

Test plan

• Run agent-builder skill and verify it loads config correctly
• Run workflow-builder skill and verify it loads config correctly
• Verify greeting uses {user_name} from config
• Verify output goes to {bmad_builder_output_folder}

Summary by CodeRabbit

• Documentation
  • Updated Agent Builder and Workflow Builder skill configuration to load settings from YAML file instead of dynamic skill initialization. Settings now explicitly define user name, communication language, and output folder paths.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: dd817cb0-11b2-46d3-8658-c5742430eb63

📥 Commits

Reviewing files that changed from the base of the PR and between 4a79058 and 814c96a.

📒 Files selected for processing (2)

• src/skills/bmad-agent-builder/SKILL.md
• src/skills/bmad-workflow-builder/SKILL.md

---

Walkthrough

Both skill documentation files replace the previous initialization method that invoked the bmad-init skill with a direct configuration loading approach from {project-root}/_bmad/bmb/config.yaml, explicitly mapping variables for user name, communication language, and output routing.

Changes

Cohort / File(s)	Summary
Configuration Loading Updates src/skills/bmad-agent-builder/SKILL.md, src/skills/bmad-workflow-builder/SKILL.md	Updated "On Activation" steps to replace bmad-init skill invocation with direct YAML config loading. Now explicitly maps {user_name}, {communication_language}, {bmad_builder_output_folder}, and {bmad_builder_reports} from config file, removing previous fallback inference logic.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Poem

🐰 A simpler path through config files we hop,
No skill-chaining loops—direct loading won't stop!
From YAML we gather our names and our ways,
Clean initialization for all our bright days! 🌿

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/metadata-finalization

📝 Coding Plan

• Generate coding plan for human review comments

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[augmentcode]

🤖 Augment PR Summary

Summary: Updates the agent/workflow builder skills to load module configuration directly from {project-root}/_bmad/bmb/config.yaml instead of invoking bmad-init.
Why: Removes the external init-skill dependency and makes the expected config variables explicit for greetings, language, and output paths.

_{🤖 Was this summary useful? React with 👍 or 👎}

[augmentcode]

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

[augmentcode]

Since bmad-init is no longer invoked, what should the skill do if {project-root}/_bmad/bmb/config.yaml is missing/unreadable or missing expected keys (e.g., first run)? Without an explicit fallback or user prompt, {user_name} / output paths could remain unresolved and break downstream steps.

Severity: medium

Other Locations

• src/skills/bmad-workflow-builder/SKILL.md:29

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}

[augmentcode]

This skill’s prompts reference {document_output_language} (e.g., quality-optimizer.md), but activation no longer mentions loading/resolving it from config. Consider ensuring it’s loaded alongside {communication_language} so report/document generation doesn’t end up with an unresolved placeholder.

Severity: medium

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}