minor dev agent updates
This commit is contained in:
Brian Madison
@@ -17,7 +17,7 @@ Specialized AI agents for different development roles:
|
||||
- **Architect** - Technical architecture and design
|
||||
- **SM** (Scrum Master) - Sprint and story management
|
||||
- **DEV** (Developer) - Code implementation
|
||||
- **SR** (Senior Reviewer) - Code review and quality
|
||||
- **TEA** (Test Architect) - Test Architect
|
||||
- **UX** - User experience design
|
||||
- And more specialized roles
|
||||
|
||||
@@ -65,17 +65,9 @@ Test architecture and quality assurance components.
|
||||
## Quick Start
|
||||
|
||||
```bash
|
||||
# Run a planning workflow
|
||||
bmad pm plan-project
|
||||
|
||||
# Create a new story
|
||||
bmad sm create-story
|
||||
|
||||
# Run development workflow
|
||||
bmad dev develop
|
||||
|
||||
# Review implementation
|
||||
bmad sr review-story
|
||||
# Load the PM agent - either via slash command or drag and drop or @ the agent file.
|
||||
# Once loaded, the agent should greet you and offer a menu of options. You can enter:
|
||||
`*plan-project`
|
||||
```
|
||||
|
||||
## Key Concepts
|
||||
|
||||
@@ -16,19 +16,20 @@ agent:
|
||||
- I treat the Story Context XML as the single source of truth, trusting it over any training priors while refusing to invent solutions when information is missing.
|
||||
- My implementation philosophy prioritizes reusing existing interfaces and artifacts over rebuilding from scratch, ensuring every change maps directly to specific acceptance criteria and tasks.
|
||||
- I operate strictly within a human-in-the-loop workflow, only proceeding when stories bear explicit approval, maintaining traceability and preventing scope drift through disciplined adherence to defined requirements.
|
||||
- I implement and execute tests ensuring complete coverage of all acceptance criteria, I do not cheat or lie about tests, I always run tests without exception, and I only declare a story complete when all tests pass 100%.
|
||||
|
||||
critical_actions:
|
||||
- "DO NOT start implementation until a story is loaded and Status == Approved"
|
||||
- "When a story is loaded, READ the entire story markdown"
|
||||
- "Locate 'Dev Agent Record' → 'Context Reference' and READ the referenced Story Context file(s). If none present, HALT and ask user to run @spec-context → *story-context"
|
||||
- "Pin the loaded Story Context into active memory for the whole session; treat it as AUTHORITATIVE over any model priors"
|
||||
- "For *develop (Dev Story workflow), execute continuously without pausing for review or 'milestones'. Only halt for explicit blocker conditions (e.g., required approvals) or when the story is truly complete (all ACs satisfied and all tasks checked)."
|
||||
- "For *develop (Dev Story workflow), execute continuously without pausing for review or 'milestones'. Only halt for explicit blocker conditions (e.g., required approvals) or when the story is truly complete (all ACs satisfied, all tasks checked, all tests executed and passing 100%)."
|
||||
|
||||
menu:
|
||||
- trigger: develop
|
||||
workflow: "{project-root}/bmad/bmm/workflows/4-implementation/dev-story/workflow.yaml"
|
||||
description: Execute Dev Story workflow (implements tasks, tests, validates, updates story)
|
||||
description: "Execute Dev Story workflow, implementing tasks and tests, or performing updates to the story"
|
||||
|
||||
- trigger: review
|
||||
workflow: "{project-root}/bmad/bmm/workflows/4-implementation/review-story/workflow.yaml"
|
||||
description: Perform Senior Developer Review on a story flagged Ready for Review (loads context/tech-spec, checks ACs/tests/architecture/security, appends review notes)
|
||||
description: "Perform a thorough clean context review on a story flagged Ready for Review, and appends review notes to story file"
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
<step n="1">Load persona from this current agent file (already in context)</step>
|
||||
<step n="2">🚨 IMMEDIATE ACTION REQUIRED - BEFORE ANY OUTPUT:
|
||||
- Use Read tool to load {project-root}/bmad/{{module}}/config.yaml NOW
|
||||
- Load and read {project-root}/bmad/{{module}}/config.yaml NOW
|
||||
- Store ALL fields as session variables: {user_name}, {communication_language}, {output_folder}
|
||||
- VERIFY: If config not loaded, STOP and report error to user
|
||||
- DO NOT PROCEED to step 3 until config is successfully loaded and variables stored</step>
|
||||
|
||||
@@ -51,13 +51,15 @@ class ActivationBuilder {
|
||||
// 2. Build menu handlers section with dynamic handlers
|
||||
const menuHandlers = await this.loadFragment('menu-handlers.xml');
|
||||
|
||||
// Build extract list (comma-separated list of used attributes)
|
||||
const extractList = profile.usedAttributes.join(', ');
|
||||
|
||||
// Build handlers (load only needed handlers)
|
||||
const handlers = await this.buildHandlers(profile);
|
||||
|
||||
const processedHandlers = menuHandlers.replace('{DYNAMIC_EXTRACT_LIST}', extractList).replace('{DYNAMIC_HANDLERS}', handlers);
|
||||
// Remove the extract line from the final output - it's just build metadata
|
||||
// The extract list tells us which attributes to look for during processing
|
||||
// but shouldn't appear in the final agent file
|
||||
const processedHandlers = menuHandlers
|
||||
.replace('<extract>{DYNAMIC_EXTRACT_LIST}</extract>\n', '') // Remove the entire extract line
|
||||
.replace('{DYNAMIC_HANDLERS}', handlers);
|
||||
|
||||
activation += '\n' + this.indent(processedHandlers, 2) + '\n';
|
||||
|
||||
|
||||
@@ -4,12 +4,29 @@
|
||||
|
||||
Aside from stability and bug fixes found during the alpha period - the main focus will be on the following:
|
||||
|
||||
- In Progress - Brownfield v6 integrated into the workflow.
|
||||
- In Progress - Full workflow single file tracking.
|
||||
- In Progress - Codex improvements.
|
||||
- Advanced Elicitation is not working well with codex
|
||||
- Brainstorming is somewhat ok with codex, but could be improved
|
||||
- Validate Gemini CLI - is it able to function at all for any workflows?
|
||||
- BoMB Tooling included with module install
|
||||
- Teat Architect better integration into workflows
|
||||
- Document new agent workflows.
|
||||
- need to segregate game dev workflows and potentially add as an installation choice
|
||||
- the workflow runner needs to become a series of targeted workflow injections at install time so workflows can be run directly without the bloated intermediary.
|
||||
- All project levels (0 through 4) manual flows validated through workflow phase 1-4 for greenfield and brownfield
|
||||
- NPX installer
|
||||
- github pipelines, branch protection, vulnerability scanners
|
||||
- subagent injections
|
||||
- bmm existing project scanning and integration with workflow phase 0-4 improvements
|
||||
- Additional custom sections for architecture project types
|
||||
- DONE: Single Agent web bundler finalized - run `npm run bundle'
|
||||
- DONE: 4->v6 upgrade installer fixed.
|
||||
- DONE: v6->v6 updates will no longer remove custom content. so if you have a new agent you created for example anywhere under the bmad folder, updates will no longer remove them.
|
||||
- DONE: if you modify an installed file and upgrade, the file will be saved as a .bak file and the installer will inform you.
|
||||
- DONE: Game Agents comms style WAY to over the top - reduced a bit.
|
||||
- need to nest subagents for better organization.
|
||||
- DONE: need to nest subagents for better organization.
|
||||
- DONE: Quick note on BMM v6 Flow
|
||||
- DONE: CC SubAgents installed to sub-folders now.
|
||||
- DONE: Qwen TOML update.
|
||||
@@ -19,24 +36,13 @@ Aside from stability and bug fixes found during the alpha period - the main focu
|
||||
- DONE: Agent improvement to loading instruction insertion and customization system overhaul
|
||||
- DONE: Stand along agents now will install to bmad/agents and are able to be compiled by the installer also
|
||||
- bmm `testarch` integrated into the BMM workflow's after aligned with the rest of bmad method flow.
|
||||
- Document new agent workflows.
|
||||
- need to segregate game dev workflows and potentially add as an installation choice
|
||||
- the workflow runner needs to become a series of targeted workflow injections at install time so workflows can be run directly without the bloated intermediary.
|
||||
- All project levels (0 through 4) manual flows validated through workflow phase 1-4
|
||||
- level 0 (simple addition or update to existing project) workflow is super streamlined from explanation of issue through code implementation
|
||||
- simple spec file -> context -> implementation
|
||||
- level 1 (simple update to existing, or a very simple oneshot tool or project)
|
||||
- NPX installer
|
||||
- github pipelines, branch protection, vulnerability scanners
|
||||
- improved subagent injections
|
||||
- bmm existing project scanning and integration with workflow phase 0-4 improvements
|
||||
- BTA Module coming soon!
|
||||
|
||||
## Needed before Beta → v0 release
|
||||
|
||||
Once the alpha is stabilized and we switch to beta, work on v4.x will freeze and the beta will merge to main. The NPX installer will still install v4 by default, but people will be able to npm install the beta version also.
|
||||
|
||||
- Orchestration tracking works consistently across all workflow phases on BMM module
|
||||
- Single Reference Architecture
|
||||
- Module repository and submission process defined
|
||||
- Final polished documentation and user guide for each module
|
||||
- Final polished documentation for overall project architecture
|
||||
@@ -49,5 +55,6 @@ Once the alpha is stabilized and we switch to beta, work on v4.x will freeze and
|
||||
- Installer offers installation of vetted community modules
|
||||
- DevOps Module
|
||||
- Security Module
|
||||
- BoMB improvements
|
||||
- Further BoMB improvements
|
||||
- 2-3 functional Reference Architecture Project Scaffolds and community contribution process defined
|
||||
-
|
||||
|
||||
Reference in New Issue
Block a user