create and edit workflow minor fixes

This commit is contained in:
Brian Madison 2025-12-27 21:35:39 +08:00
parent 2a8a4388a9
commit ea421adbf9
8 changed files with 47 additions and 66 deletions

View File

@ -90,7 +90,7 @@ Wait for clear user response (yes/no or y/n).
- Load brainstorming workflow: `{brainstormWorkflow}` - Load brainstorming workflow: `{brainstormWorkflow}`
- Pass context data: `{brainstormContext}` - Pass context data: `{brainstormContext}`
- Execute brainstorming session scoped specifically to brainstorming a new agent. - Execute brainstorming session scoped specifically utilizing the brainstormContext to guide the scope and outcome
- Capture all brainstorming output for next step - Capture all brainstorming output for next step
- Return to this step after brainstorming completes - Return to this step after brainstorming completes

View File

@ -1,3 +1,5 @@
--- ---
stepsCompleted: [] stepsCompleted: []
--- ---
# Agent Design and Build Plan

View File

@ -28,28 +28,22 @@ Domain-specific agents with persistent memory, sidecar files, and restricted acc
```yaml ```yaml
agent: agent:
metadata: metadata:
id: _bmad/agents/{agent-name}/{agent-name}.md
name: 'Persona Name' name: 'Persona Name'
title: 'Agent Title' title: 'Agent Title'
icon: 'emoji' icon: 'emoji'
type: 'expert' module: stand-alone # or module name
persona: persona:
role: 'Domain Expert with specialized capability' role: |
First-person description of primary function (1-2 sentences)
identity: | identity: |
Background and expertise in first-person voice. Background, experience, specializations in first-person (2-3 sentences)
{{#if user_preference}}
Customization based on install_config.
{{/if}}
communication_style: | communication_style: |
{{#if tone_style == "gentle"}} 1-2 short sentence describe how the agent speaks and communicates
Gentle and supportive communication...
{{/if}}
{{#if tone_style == "direct"}}
Direct and efficient communication...
{{/if}}
I reference past conversations naturally.
principles: principles:
- Core belief about the domain - Core belief about the domain
@ -89,29 +83,17 @@ agent:
Reference stored information naturally. Reference stored information naturally.
menu: menu:
- trigger: action1 - trigger: MF or fuzzy match on main function
action: '#main-function' action: '#main-function'
description: 'Primary agent function' description: '[MF] Main agent function'
- trigger: remember - trigger: SM or fuzzy match on save-memory
action: 'Update ./{agent-name}-sidecar/memories.md with session insights' action: 'Update ./{agent-name}-sidecar/memories.md with session insights'
description: 'Save what we discussed today' description: '[SM] Save memory what we discussed today'
- trigger: insight - trigger: RI or fuzzy match on record-insight
action: 'Document breakthrough in ./{agent-name}-sidecar/breakthroughs.md' action: 'Document breakthrough in ./{agent-name}-sidecar/breakthroughs.md'
description: 'Record a significant insight' description: '[RI] Record a significant insight'
- multi: "[DF] Do Foo or start [CH] Chat with expert"
triggers:
- do-foo
- input: [DF] or fuzzy match on do foo
- action: '#main-action'
- data: what is being discussed or suggested with the command, along with custom party custom agents if specified
- type: action
- expert-chat:
- input: [CH] or fuzzy match validate agent
- action: agent responds as expert based on its persona to converse
- type: action
install_config: install_config:
compile_time_only: true compile_time_only: true

View File

@ -18,7 +18,7 @@ agent:
name: 'Persona Name' name: 'Persona Name'
title: 'Agent Title' title: 'Agent Title'
icon: 'emoji' icon: 'emoji'
type: simple module: stand-alone # or module name
persona: persona:
role: | role: |
@ -26,17 +26,9 @@ agent:
identity: | identity: |
Background, experience, specializations in first-person (2-5 sentences) Background, experience, specializations in first-person (2-5 sentences)
{{#if custom_variable}}
Conditional identity text based on install_config
{{/if}}
communication_style: | communication_style: |
{{#if style_choice == "professional"}} 1-2 short sentence describe how the agent speaks and communicates
Professional and systematic approach...
{{/if}}
{{#if style_choice == "casual"}}
Friendly and approachable tone...
{{/if}}
principles: principles:
- Core belief or methodology - Core belief or methodology
@ -58,26 +50,18 @@ agent:
3. Final step 3. Final step
</process> </process>
- id: another-action - id: foo-bar
content: | content: |
Another reusable prompt template Another reusable prompt template
menu: menu:
- trigger: inline - trigger: FB or fuzzy match on foo-bar
action: 'Direct inline prompt text' action: '#foo-bar'
description: 'Execute inline action' description: '[FB] Foo Bar inline action'
- multi: "[DF] Do Foo or start [CH] Chat with expert" - trigger: IA or fuzzy match on main-action
triggers: action: '#main-action'
- do-foo description: '[IA] Execute inline action'
- input: [DF] or fuzzy match on do foo
- action: '#main-action'
- data: what is being discussed or suggested with the command, along with custom party custom agents if specified
- type: action
- expert-chat:
- input: [CH] or fuzzy match validate agent
- action: agent responds as expert based on its persona to converse
- type: action
install_config: install_config:
compile_time_only: true compile_time_only: true

View File

@ -96,9 +96,9 @@ Read: {project-root}/_bmad/bmb/docs/workflows/templates/step-1b-template.md
This shows the continuation step pattern for workflows that might take multiple sessions. This shows the continuation step pattern for workflows that might take multiple sessions.
Based on the approved plan, collaboratively design: Based on the approved plan, collaboratively design the info to answer the following for the build plan:
- How many major steps does this workflow need? (Recommend 3-7) - How many major steps does this workflow need?
- What is the goal of each step? - What is the goal of each step?
- Which steps are optional vs required? - Which steps are optional vs required?
- Should any steps repeat or loop? - Should any steps repeat or loop?

View File

@ -72,7 +72,7 @@ To generate all the workflow files (workflow.md, step files, templates, and supp
- When building continuable step-01-init.md files, use template `{project-root}/_bmad/bmb/docs/workflows/templates/step-01-init-continuable-template.md` - When building continuable step-01-init.md files, use template `{project-root}/_bmad/bmb/docs/workflows/templates/step-01-init-continuable-template.md`
- When building continuation steps, use template `{project-root}/_bmad/bmb/docs/workflows/templates/step-1b-template.md` - When building continuation steps, use template `{project-root}/_bmad/bmb/docs/workflows/templates/step-1b-template.md`
- When building the main workflow.md file, you must follow template `{project-root}/_bmad/bmb/docs/workflows/templates/workflow-template.md` - When building the main workflow.md file, you must follow template `{project-root}/_bmad/bmb/docs/workflows/templates/workflow-template.md`
- Example step files from {project-root}/_bmad/bmb/reference/workflows/meal-prep-nutrition/workflow.md for patterns - Example step files from {project-root}/_bmad/bmb/reference/workflows/meal-prep-nutrition/workflow.md for patterns - this is an idealized workflow so all files can give good insight into format and structure to be followed
## FILE GENERATION SEQUENCE: ## FILE GENERATION SEQUENCE:
@ -182,6 +182,20 @@ For document workflows:
- Include all variables from design - Include all variables from design
- Ensure variable naming consistency - Ensure variable naming consistency
Remember that the output format design we aligned on chose one of the following - and what it means practically when creating the workflow steps:
1. **Strict Template** - Must follow exact format with specific fields
1. This is similar to the example where there are multiple template fragements that are specific with all fields to be in the final output.
2. generally there will be 1 fragment to a step to complete in the overall template.
2. **Structured** - Required sections but flexible within each
1. Usually there will just be one template file - and in this mode it lists out all the section headings (generally level 2 sections in the md) with a handlebars style placeholder for each section.
2. Step files responsible for a specific section will upon user Continue of that step ensure output is written to the templates proper section
3. **Semi-structured** - Core sections plus optional additions
1. Similar to the prior 2, but not all sections or content are listed in the template, some steps might offer various paths or options to go to different steps (or variance within a step) that can determine what sections end up in the final document
4. **Free-form** - Content-driven with minimal structure
1. These are the easiest and most flexible. The single template usually only has the front matter fence with a stepsCompleted array and maybe some other fields, and outside of the front matter just the level 1 doc title
2. With free form, any step that could produce content just appends to the end of the document, so its progressively build in the order of ste[s completed.
3. Its good to have in this type of workflow a final polish output doc type step that cohesively can update the doc built up in this progressive manner, improving flow, reducing duplication, and ensure all information is aligned and where it belongs.
### 6. Generate Supporting Files ### 6. Generate Supporting Files
Based on design requirements: Based on design requirements:

View File

@ -127,7 +127,7 @@ Based on what the user wants to edit:
#### Compliance Analysis #### Compliance Analysis
Load reference documentation as needed: Load reference documentation to understand what ideal workflow files sound be when doing the review:
- `{project-root}/_bmad/bmb/docs/workflows/architecture.md` - `{project-root}/_bmad/bmb/docs/workflows/architecture.md`
- `{project-root}/_bmad/bmb/docs/workflows/templates/step-template.md` - `{project-root}/_bmad/bmb/docs/workflows/templates/step-template.md`
@ -135,10 +135,9 @@ Load reference documentation as needed:
Check against best practices: Check against best practices:
- Step file size and structure - Step file size and structure (each step file 80-250 lines)
- Menu handling implementation - Menu handling implementation (every menu item has a handler, and continue will only proceed after writes to output if applicable have completed)
- Frontmatter variable usage - Frontmatter variable usage - no unused variables in the specific step front matter, and all files referenced in the file are done through a variable in the front matter
- Path reference consistency
### 5. Present Analysis Findings ### 5. Present Analysis Findings

View File

@ -51,7 +51,7 @@ This uses **step-file architecture** for disciplined execution:
Load and read full config from {project-root}/_bmad/bmb/config.yaml and resolve: Load and read full config from {project-root}/_bmad/bmb/config.yaml and resolve:
- `project_name`, `output_folder`, `user_name`, `communication_language`, `document_output_language` - `project_name`, `output_folder`, `user_name`, `communication_language`, `document_output_language`, `bmb_creations_output_folder`
### 2. First Step EXECUTION ### 2. First Step EXECUTION