Standalone instruction file that patches existing BMAD installations
to use explicit {interaction_style} session variable lookup with
open-mode default fallback.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Update module.yaml default to "open" and reorder options
- Change fallback in all 40 RULE directives from "default to structured
mode" to "default to open mode"
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Apply explicit {interaction_style} variable reference to handler-action,
handler-exec, handler-workflow, agent-command-header, and workflow.xml
- Remove old duplicate RULE lines left over from merge conflict resolution
in 10 template files (default-task/tool, rovodev, opencode-workflow*,
gemini-agent/task/tool/workflow/workflow-yaml)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Merges feature/platform-specific-structured-interaction which adds:
- Per-IDE template variants (claude, opencode, gemini, kiro, etc.)
- Explicit {interaction_style} session variable lookup in all RULE directives
- Fallback default to structured mode when variable is undetermined
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Change all 34 RULE directives from bare "When interaction_style is"
to explicit "Check your stored {interaction_style} session variable
from activation step 2" with a fallback default to structured mode.
Addresses cynical review finding that LLMs struggle with implicit
natural-language conditionals and need explicit variable lookup
instructions plus failure mode handling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Matches the established pattern where all session variables use
{braces} when referenced (like {communication_language} and
{user_name}). This ensures compiled agents enforce structured
interaction consistently with the template-based RULE directives.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Standardize all 27 template RULE lines to use "EVERY question"
phrasing. The free-form escape clause prevents over-restriction
while EVERY strengthens the default structured behavior.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Create claude-specific template set (agent, task, tool, workflow,
workflow-yaml) that explicitly references the `AskUserQuestion` tool,
matching how gemini/kiro/opencode each have their own templates with
IDE-specific tool names.
- Change Claude Code template_type from "default" to "claude" in
platform-codes.yaml
- Replace symlinks with real template files
- Create new claude-task.md and claude-tool.md
- default-* templates remain as generic fallbacks for cursor, auggie,
and other IDEs using numbered-list wording
- Fix default-tool.md RULE placement (move before execution instructions)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace emphatic "EVERY...MUST...Do NOT...No exceptions" phrasing with
clearer natural language that includes an explicit free-form override.
Research shows single clear sentences with conditional logic outperform
stacked absolute directives for LLM compliance.
New wording: "questions to the user MUST use structured options with a
numbered list unless the workflow or user explicitly requests free-form
input."
Also extracts RULE from step 4 in gemini-task.toml to standalone.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Replace Claude-specific AskUserQuestion in default-* templates with generic
numbered-list wording since these are universal fallbacks for all IDEs
- Fix hardcoded _bmad paths to use {{bmadFolderName}} in antigravity, rovodev,
trae, windsurf-workflow, and default-agent templates
- Extract RULE from numbered steps to standalone directive in gemini and
opencode workflow/tool templates
- Standardize RULE wording: add "Do NOT ask open-ended inline questions" to
all templates missing it for consistency
- Move RULE before execution instructions in default-task and rovodev templates
- Add {document_output_language} to activation-steps session variables
- Add interaction_style confirmation to gemini-agent preflight checklist
- Document all config settings including interaction_style in install guide
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Method-driven review that exhaustively walks branching paths and
boundary conditions, reporting only unhandled gaps. Orthogonal to
the attitude-driven adversarial review for complementary coverage.
Isolated test branch with just IDE template rules and module.yaml config.
No core agent-component or workflow.xml changes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds a new `interaction_style` config option (structured/open) to core
module.yaml so users can choose how agents ask questions. When set to
"structured" (default), agents present numbered option lists and wait
for selection. When "open", agents ask questions naturally.
Changes:
- src/core/module.yaml: new interaction_style config with single-select
- activation-rules.txt: conditional rule based on {interaction_style}
- activation-steps.txt: include {interaction_style} in stored variables
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add all changes since v6.0.2 including new root-cause-analysis skill,
installer fixes, workflow fixes, and documentation updates.
Also standardize quotes in remaining task/workflow files.
New internal skill that analyzes bug-fix commits or PRs and produces
a structured Root Cause Analysis report with pyramid communication.
Co-authored-by: Brian <bmadcode@gmail.com>
* fix(installer): refuse install when ancestor dir has BMAD commands
Claude Code inherits slash commands from parent directories, so
installing into a nested project when a parent already has .claude/commands
with bmad-* files causes duplicate entries in the autocomplete.
Add ancestor_conflict_check flag (enabled for claude-code) that walks
up the directory tree before install. If BMAD files are found in an
ancestor target_dir, the installer refuses with an actionable error.
Also fix IdeManager.setup() to propagate handler success status instead
of unconditionally returning success: true.
* Address code review feedback from CodeRabbit and Augment
- Move "Setting up..." log after conflict check so it only shows when
install will proceed
- Fix rm command: add -rf flags and correct quoting for glob outside quotes
- Improve error wording: "ancestor installation" instead of misleading
"ancestor directory"
- Use case-insensitive startsWith for bmad file detection (macOS/Windows)
- Document ancestor_conflict_check in the installer config schema
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(installer): resolve symlinks before ancestor conflict walk
Use fs.realpath() instead of path.resolve() so the ancestor directory
walk follows the physical filesystem path, not the logical symlink path.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix(opencode): use mode: all in agent template, remove name frontmatter, fix directory names
- Replace name: '{{name}}' with mode: all in opencode-agent.md
mode: all enables both Tab-key agent switching in the TUI and @subagent
invocation via the Task tool (mode: primary blocked subagent use)
- Remove name: '{{name}}' from opencode-task/tool/workflow/workflow-yaml templates
OpenCode derives command name from filename, not from a name frontmatter field;
the bare {{name}} value was overriding the bmad- prefixed filename causing
name collisions with built-in OpenCode commands (fixes#1762)
- Fix deprecated singular directory names in platform-codes.yaml:
.opencode/agent -> .opencode/agents, .opencode/command -> .opencode/commands
- Add legacy_targets migration: cleanup() now removes stale bmad-* files from
old singular directories on reinstall so existing users don't get duplicates
- Fix removeEmptyParents to continue walking up to parent when starting dir is
already absent instead of breaking early
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(opencode): address code review findings for cleanup and schema docs
- Add project boundary guard to removeEmptyParents() using path.resolve
and startsWith check to prevent traversal outside projectDir (Augment)
- Fix JSDoc: "Recursively remove" -> "Walk up ancestor directories"
- Add user-visible migration log message when processing legacy_targets
- Document legacy_targets field in Installer Config Schema comment block
in platform-codes.yaml (CodeRabbit + Augment)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(opencode): improve removeEmptyParents error handling and loop clarity
- Distinguish recoverable errors (ENOTEMPTY, ENOENT) from fatal errors in
removeEmptyParents() catch block — skip level and continue upward on
TOCTOU races or concurrent removal, break only on fatal errors (EACCES)
- Add comment clarifying loop invariant for missing-path continue branch
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Brian <bmadcode@gmail.com>
Update the BMAD acronym expansion from "Breakthrough Method of
Agile AI Driven Development" to "Build More Architect Dreams"
across README, docs homepage, and package.json description.
Co-authored-by: Brian <bmadcode@gmail.com>
* fix(installer): remove double-escaping of quotes in CSV manifest pipeline
cleanForCSV() pre-escaped " to "" before storing in memory, then
escapeCsv() escaped again at CSV write time. After csv-parse round-trip
(which only un-escapes once), descriptions retained doubled quotes
instead of originals, corrupting generated output files.
Fix: remove the redundant quote escaping from cleanForCSV() since
escapeCsv() already handles CSV quoting correctly at write time.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(installer): use single quotes for description in Gemini workflow templates
Replace triple-quoted """{{description}}""" with single-quoted '{{description}}'
to avoid TOML escaping issues in Gemini workflow templates.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix(workflows): remove ambiguous with-argument from help task chaining
The "with argument" clause in 7 workflow completion steps caused LLMs
to interpret "Read fully and follow: help.md with argument X" as a
skill/function invocation rather than a file-read instruction. Drop the
clause entirely — help.md already infers the completed workflow from
the preceding "[Workflow] complete." text in conversation context.
Closes#1637
* fix(workflows): correct broken qa/automate file references
The QA workflow was renamed to qa-generate-e2e-tests but three files
still referenced the old qa/automate path, breaking CI file-ref
validation.
* fix(test): update QA agent test to match renamed workflow path
The workflow path changed from qa/automate to qa-generate-e2e-tests
but the installation component test was not updated to match.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Brian <bmadcode@gmail.com>
- Add comprehensive public roadmap with In Progress, Getting Started, and Community sections
- Add roadmap callout to README.md and docs welcome page
- Add roadmap-specific card grid styles with equal-height cards
- Fix rehype-markdown-links plugin to detect docs directory in various project structures
- Rename roadmap.md to roadmap.mdx for proper Astro/Starlight rendering
- Update doc link validator to support .mdx files
This change removes the disable-model-invocation flag from all IDE installer
templates. By allowing model invocation, bmad help can now properly invoke
suggested workflows as direct skill calls, improving the user experience by
enabling automatic workflow execution when desired.
- Standardize all workflow descriptions to follow format: [short description]. Use when the user says 'explicit action phrase' or 'another phrase'
- Remove verbose descriptions in favor of concise summaries with explicit trigger phrases
- Use max 2 phrases per workflow to minimize context and false positives
- Phrases are explicit actions (e.g., "lets create", "run X") not questions
- No slash commands in descriptions - users invoke via /name directly
- Rename qa/automate to qa-generate-e2e-tests for clarity
- Update various core tasks and workflows
* feat: migrate Codex installer from .codex/prompts to .agents/skills format
Switch CodexSetup to write BMAD artifacts as Agent Skills (agentskills.io
format) in .agents/skills/<name>/SKILL.md instead of flat files in
.codex/prompts. Remove global/project location prompt. Add legacy cleanup
of old .codex/prompts directories during install and uninstall.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* ignore .agents for codex
* fix: normalize line endings and use platform-native EOL in SKILL.md output
Normalize all content to LF in transformToSkillFormat, then convert to
os.EOL (CRLF on Windows, LF on Linux/macOS) before writing SKILL.md
files in both writeSkillArtifacts and installCustomAgentLauncher.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* ignore .agents directory
* use description from metadata in custom agent launcher
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Move Raven PR review and file-ref audit from tools/ into
.claude/skills/ as proper bmad-os skills with SKILL.md + prompts/
instructions.md split pattern. Strip XML tags from Raven content,
promote sections to H2 headings. Delete originals from tools/.
Co-authored-by: Brian <bmadcode@gmail.com>
Update 9 files that referenced the old path
bmm/workflows/3-solutioning/architecture/steps/ to the correct
bmm/workflows/3-solutioning/create-architecture/steps/ after the
directory was renamed.
Fixes#1625
Co-authored-by: Junie <junie@jetbrains.com>
tools/audit-file-refs.md — a repeatable prompt that spawns parallel
Haiku subagents to semantically audit new-format source files for
non-conforming file references. Includes a self-check that verifies
all files are accounted for before producing the final report.
Replaces the planned regex extension (Item 1 of the master plan) with
an approach that can handle the full surface area of reference patterns
without exhaustive pattern enumeration.
Also excludes .junie/ from Prettier checks (IDE integration folder,
user-specific, not in repo).
Refs #1718
Co-authored-by: Brian <bmadcode@gmail.com>
Add CodeBuddy (Tencent Cloud Code Assistant) as a supported IDE platform:
- Add platform config to tools/platform-codes.yaml
- Add installer config to tools/cli/installers/lib/ide/platform-codes.yaml
- Add .codebuddy to .gitignore
CodeBuddy uses the default template type with target directory .codebuddy/commands
Co-authored-by: wison <wisonlin@tencent.com>
Co-authored-by: Brian <bmadcode@gmail.com>
- create-architecture/workflow.md: fix installed_path dir name
from 'architecture' to 'create-architecture'
- create-story/checklist.md: fix 2 refs from validate-workflow.xml
to workflow.xml (file does not exist with validate- prefix)
- package.json: add --strict to validate:refs so broken references
fail CI instead of logging warnings and exiting 0
Co-authored-by: Brian <bmadcode@gmail.com>
* fix(installer): add custom Rovo Dev installer with prompts.yml generation
Rovo Dev CLI requires a .rovodev/prompts.yml manifest to register prompts
for /prompts access. The config-driven installer was writing .md files but
never generating this manifest, so /prompts showed nothing.
- Create custom rovodev.js installer extending BaseIdeSetup
- Generate prompts.yml indexing all written workflow files
- Merge with existing user entries (only touch bmad- prefixed entries)
- Remove stale rovo entry from tools/platform-codes.yaml
Closes#1466
* fix(installer): prefix prompts.yml descriptions with entry name
The /prompts list in Rovo Dev only shows descriptions, making it hard
to identify entries. Prefix each description with the bmad entry name
so users see e.g. "bmad-bmm-create-prd - PRD workflow..." instead of
just the description text.
* refactor(installer): address review findings in Rovo Dev installer
- Hoist toDashPath import to module top level
- Extract _collectPromptEntries helper replacing 3 duplicated loops
- Remove unused detectionPaths (detect() is overridden)
- Guard generatePromptsYml when writtenFiles is empty
- Align cleanup() with detect() predicate (remove any bmad-*, not just .md)
- Use BaseIdeSetup abstractions (this.pathExists/readFile/writeFile) in cleanup()
- Update loadHandlers() JSDoc to include rovodev.js
---------
Co-authored-by: Brian <bmadcode@gmail.com>
* fix: replace bare _bmad/ backtick refs with {project-root}/_bmad/
Closes#1718 (partial — bare _bmad/ category only)
* fix: replace relative step refs with {project-root}/_bmad/ paths
Converts all ./step-XX.md, step-XX.md, and steps/step-XX.md
backtick references in new-format workflow and step files to
full {project-root}/_bmad/... paths.
Refs #1718
* fix: correct create-architecture installed path and remaining relative refs
- replace create-architecture with architecture in all step path refs
to match workflow.md installed_path definition
- convert ../data/ relative refs in create-prd step-05/06/11 frontmatter
- fix stale nextStepFile example in create-prd step-01b-continue
- fix bare step-01-init.md ref in create-architecture step-01b-continue
Fixes#1718
* fix: convert remaining relative data refs and fix stale examples in continue steps
* fix: inline quick-spec step navigation paths, remove frontmatter tokens
Replace {nextStepFile} and {skipToStepFile} frontmatter tokens with
explicit {project-root}/_bmad/ paths in all quick-spec step files.
These are LLM prompts, not config files -- inline paths are clearer
and carry semantic information without indirection. Also standardize
wording from "Load" to "Read fully and follow:" for consistency.
Also add .junie/ to .prettierignore to fix unrelated CI noise.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Brian <bmadcode@gmail.com>
The 4-backtick markdown fence closed prematurely at line 235,
orphaning template content and causing a stray 3-backtick fence
to swallow sections 9-13 into an unclosed code block.
Windsurf is no longer a preferred IDE. Only Claude Code and Cursor
are now recommended. Windsurf remains a supported platform (installer
still works, templates stay, reference tables stay).
- Set preferred: false in both platform-codes.yaml files
- Move windsurf entry to "Other IDEs" section in tools/platform-codes.yaml
- Fix codex.js hardcoding preferred: true in constructor
- Remove stale "3 preferred tools" count from ui.js JSDoc
- Update docs to list only Claude Code and Cursor as recommended
- Update docs/index.md popular tools to Claude Code, Cursor, Codex CLI
Sidharth Arora
Alex Verkhovsky
Chris Chen
Brian Madison
Ssabae
Davor Racic
PinkyD
Wendy Smoak
wison
Daniel Bates
H Paul Hammann