Issue Type Schema

Issue Type Schema

Schema definition for an issuetype template

Schema Information

Required Fields

Properties

Property Type Required Description
key string Yes Unique issuetype key (e.g., FEAT, FIX, SPIKE, INV, TASK)
name string Yes Display name of the template type
description string Yes Brief description of when to use this template
mode → ExecutionMode Yes Whether this issuetype runs autonomously or requires human pairing
glyph string Yes Glyph character displayed in UI for this issuetype
color string | null No Optional color for glyph display in TUI
project_required boolean No Whether a project must be specified for this issuetype
fields array Yes Field definitions for this template
steps array Yes Lifecycle steps for completing this ticket type
prompt string | null No Optional prompt for work launching (interpolated with handlebars)
agent_prompt string | null No Prompt for generating this issue type’s operator agent via claude -p
agent string | null No Default delegator name for this issuetype (overridden by step.agent)

key

name

description

mode

glyph

color

project_required

fields

steps

prompt

agent_prompt

agent

Definitions

Definition: ExecutionMode

Execution mode for an issuetype

Definition: FieldSchema

Schema definition for a single field in a template

Property Type Required Description
name string Yes Field identifier (matches handlebar variable name)
description string Yes Help text for the field
type → FieldType Yes Type of the field
required boolean No Whether this field must be filled
default string | null No Default value if any
auto object No Auto-generation strategy for this field
options array No Options for enum fields
placeholder string | null No Placeholder text shown in template
max_length integer | null No Maximum length for string fields
display_order integer | null No Display order in form (lower = first)
user_editable boolean No Whether the user can edit this field (false for auto-generated)

Definition: FieldType

Types of fields supported in template schemas

Definition: AutoGenStrategy

Auto-generation strategies for fields

Definition: StepSchema

Schema definition for a lifecycle step

Property Type Required Description
name string Yes Step identifier (lowercase)
display_name string | null No Human-readable step name
type → StepTypeTag No Step type discriminator (defaults to “task” for backward compatibility)
outputs array Yes Types of outputs this step produces
prompt string Yes Initial prompt template for the Claude agent
review_type → ReviewType No Type of review required for this step (none, plan, visual, pr)
visual_config object No Configuration for visual review (required when review_type is “visual”)
on_reject object No What to do if step output is rejected
next_step string | null No Name of the next step (None for final step)
allowed_tools array No Claude Code tools allowed in this step
agent string | null No Optional agent (delegator) name for this step (overrides ticket’s default agent)
permissions object No Provider-agnostic permissions for this step
cli_args object No Arbitrary CLI arguments per provider
permission_mode → PermissionMode No Preferred LLM permission mode for this step
jsonSchema object No Inline JSON schema for structured output (Claude-specific)
jsonSchemaFile string | null No Path to JSON schema file for structured output (Claude-specific)
artifact_patterns array No File glob patterns in the worktree that signal this step is complete
classifier_config object No Configuration for classifier steps (required when type=classifier)
rag_config object No Configuration for RAG steps (required when type=rag)
delegator_config object No Configuration for delegator steps (required when type=delegator)
mcp_config object No Configuration for MCP steps (required when type=mcp)
multi_model_config object No Configuration for multi-model steps (required when type=multi_model)
multi_prompt_config object No Configuration for multi-prompt steps (required when type=multi_prompt)
matrixed_config object No Configuration for matrixed steps (required when type=matrixed)

Definition: StepTypeTag

Discriminator tag for step types

Definition: StepOutput

Types of outputs a step can produce

Definition: ReviewType

Type of review required for a step

Definition: VisualReviewConfig

Configuration for visual review steps

Property Type Required Description
url string Yes URL to open for visual check (supports handlebars templates)
startup_command string | null No Optional startup command (e.g., dev server) to run before opening browser
startup_timeout_secs integer | null No Timeout in seconds for server startup (default: 30)

Definition: OnReject

Action to take when a step is rejected

Property Type Required Description
goto_step string Yes Step name to return to on rejection
prompt string Yes Prompt to use when restarting after rejection

Definition: StepPermissions

Complete permission set for a step (as defined in issuetype schema)

Property Type Required Description
tools → ToolPermissions No Tool-level allow/deny lists
directories → DirectoryPermissions No Directory-level allow/deny lists
mcp_servers → McpServerPermissions No MCP server enable/disable configuration
custom_flags → CustomFlags No Per-provider custom configuration flags

Definition: ToolPermissions

Tool-level permissions (allow/deny lists)

Property Type Required Description
allow array No Tools/patterns to allow
deny array No Tools/patterns to deny

Definition: ToolPattern

Provider-agnostic tool pattern

Property Type Required Description
tool string Yes Tool name: Read, Write, Edit, Bash, Glob, Grep, WebFetch, etc.
pattern string | null No Optional pattern for tool arguments (e.g., “cargo test:*” for Bash)

Definition: DirectoryPermissions

Directory-level permissions

Property Type Required Description
allow array No Additional directories to allow access to (glob patterns)
deny array No Directories to deny access to (glob patterns)

Definition: McpServerPermissions

MCP server permissions (server-level enable/disable only)

Property Type Required Description
enable array No MCP servers to enable for this step
disable array No MCP servers to disable for this step

Definition: CustomFlags

Per-provider custom configuration flags

Property Type Required Description
claude object No Claude-specific configuration flags
gemini object No Gemini-specific configuration flags
codex object No Codex-specific configuration flags

Definition: ProviderCliArgs

Arbitrary CLI arguments per provider

Property Type Required Description
claude array No CLI arguments for Claude
gemini array No CLI arguments for Gemini
codex array No CLI arguments for Codex

Definition: PermissionMode

Permission mode for LLM interaction

Definition: ClassifierConfig

Configuration for classifier steps that return structured typed output

Property Type Required Description
output_type → ClassifierOutputType Yes What type of answer the classifier returns
options array | null No For enum type: the allowed options
max_length integer | null No For short_string: max character length (default 255)
agent string | null No Agent/delegator to use (overrides issuetype default)

Definition: ClassifierOutputType

Output types for classifier steps

Definition: RagConfig

Configuration for RAG (retrieval-augmented generation) steps

Property Type Required Description
sources array Yes Context sources to retrieve before running the prompt
max_context_tokens integer | null No Maximum tokens of context to inject (default: 50000)
agent string | null No Agent/delegator to use
allowed_tools array No Tools allowed for the agent

Definition: RagSource

A source of context for RAG steps

Definition: DelegatorStepConfig

Configuration for delegator steps that run with a specific model+flavor

Property Type Required Description
delegator string Yes Named delegator reference (from config.delegators)
prompt_flavor string | null No Additional prompt flavor text prepended to the step prompt
allowed_tools array No Tools allowed
permissions object No Permissions

Definition: McpStepConfig

Configuration for MCP steps that require specific MCP tools

Property Type Required Description
required_tools array Yes MCP tools that MUST be available (step fails if missing)
optional_tools array No MCP tools that SHOULD be available (warning if missing)
agent string | null No Agent/delegator to use
allowed_tools array No Tools allowed (in addition to MCP tools)

Definition: McpToolRef

Reference to a specific MCP server tool

Property Type Required Description
server string Yes MCP server name
tool string | null No Specific tool name (None = all tools from this server)

Definition: MultiModelConfig

Configuration for multi-model delegation steps (fan-out + vote)

Property Type Required Description
delegators array Yes Named delegator references (from config.delegators), minimum 2
voting_strategy → VotingStrategy Yes How to aggregate/select the final answer
share_answers boolean No Whether to share all answers with all models in the voting round
voting_prompt string | null No Prompt for the voting round (Handlebars, receives array)
voting_mode → VotingMode No How the voting round executes

Definition: VotingStrategy

Voting strategy for multi-model steps

Definition: VotingMode

How the voting round is executed in multi-model steps

Definition: MultiPromptConfig

Configuration for multi-prompt interrogation steps (N variations, select best)

Property Type Required Description
prompt_variations array Yes Prompt variations (Handlebars templates), minimum 2
selection_strategy → SelectionStrategy Yes How to select the best result
agent string | null No Agent/delegator to use for all variations
selection_prompt string | null No Prompt for the selection/review round

Definition: SelectionStrategy

Selection strategy for multi-prompt steps

Definition: MatrixedConfig

Configuration for matrixed work output steps (N x M delegators x prompts)

Property Type Required Description
delegators array Yes Named delegator references (N), minimum 2
prompt_variations array Yes Prompt variations (M) — Handlebars templates, minimum 2
output_format → MatrixedOutputFormat Yes How to organize/present the N x M output
aggregation_prompt string | null No Optional aggregation prompt (receives the full matrix of results)

Definition: MatrixedOutputFormat

Output format for matrixed steps