speck
Installation
Entrypoint
Types
Speck

speck

Speck — AI-powered spec-driven development agent.

Installation

dagger install github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5

Entrypoint

Return Type

Speck !

Arguments

Name	Type	Default Value	Description
source	Directory	-	Project source directory

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \

func (m *MyModule) Example() *dagger.Speck  {
	return dag.
			Speck()
}

@function
def example() -> dagger.Speck:
	return (
		dag.speck()
	)

@func()
example(): Speck {
	return dag
		.speck()
}

Types

Speck 🔗

AI-powered spec-driven development agent. Takes a prompt or GitHub issue and produces structured task decompositions with agent and model assignments. The output is a JSON object designed for GitHub Actions matrix consumption. Pass your project as --source and provide an agent registry via --agents to enable intelligent agent assignment.

source() 🔗

Project source directory

Return Type

Directory !

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 source

func (m *MyModule) Example() *dagger.Directory  {
	return dag.
			Speck().
			Source()
}

@function
def example() -> dagger.Directory:
	return (
		dag.speck()
		.source()
	)

@func()
example(): Directory {
	return dag
		.speck()
		.source()
}

decompose() 🔗

Run the full spec-driven development pipeline.

Executes specify -> plan -> decompose in sequence and returns a structured JSON object with tasks, agent assignments, model suggestions, and an execution plan. Designed for GitHub Actions matrix consumption.

When create_pr is enabled, each phase in the execution plan gets a pr_branch field. The GitHub Actions workflow should matrix by phase, chain tasks sequentially (impl -> test -> review), and create one PR per phase with the accumulated changes.

Return Type

String !

Arguments

Name	Type	Default Value	Description
prompt	String !	""	What to build (runs full pipeline: specify -> plan -> decompose)
issueId	Integer !	0	GitHub issue number (alternative to prompt)
repository	String !	""	GitHub repository URL (required if using issue_id)
githubToken	Secret	null	GitHub token (required if using issue_id)
agents	String !	"[]"	JSON array of available agents with name, source, specialization, capabilities
techStack	String !	""	Tech stack hints (e.g. 'Python, FastAPI, PostgreSQL')
modelFamily	String !	"claude"	Model family for suggestions: 'claude' (default), 'gemini', or 'openai'
includeTests	Boolean !	false	Generate test tasks (write_tests) after implementation tasks
includeReview	Boolean !	false	Generate review tasks (review) at the end of each phase
createPr	Boolean !	false	Add PR branch names to phases for automated PR creation (one PR per phase)

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 decompose --prompt string --issue-id integer --repository string --agents string --tech-stack string --model-family string --include-tests boolean --include-review boolean --create-pr boolean

func (m *MyModule) Example(ctx context.Context, prompt string, issueId int, repository string, agents string, techStack string, modelFamily string, includeTests bool, includeReview bool, createPr bool) string  {
	return dag.
			Speck().
			Decompose(ctx, prompt, issueId, repository, agents, techStack, modelFamily, includeTests, includeReview, createPr)
}

@function
async def example(prompt: str, issue_id: int, repository: str, agents: str, tech_stack: str, model_family: str, include_tests: bool, include_review: bool, create_pr: bool) -> str:
	return await (
		dag.speck()
		.decompose(prompt, issue_id, repository, agents, tech_stack, model_family, include_tests, include_review, create_pr)
	)

@func()
async example(prompt: string, issueId: number, repository: string, agents: string, techStack: string, modelFamily: string, includeTests: boolean, includeReview: boolean, createPr: boolean): Promise<string> {
	return dag
		.speck()
		.decompose(prompt, issueId, repository, agents, techStack, modelFamily, includeTests, includeReview, createPr)
}

decomposeFromSpec() 🔗

Decompose a pre-existing spec and plan into structured tasks.

Use this when you already have a specification and plan (e.g. from manual spec-kit usage) and want to generate the structured JSON task list with agent assignments.

Return Type

String !

Arguments

Name	Type	Default Value	Description
spec	String !	-	Feature specification markdown
plan	String !	-	Technical implementation plan markdown
agents	String !	"[]"	JSON array of available agents
modelFamily	String !	"claude"	Model family for suggestions: 'claude' (default), 'gemini', or 'openai'
includeTests	Boolean !	false	Generate test tasks (write_tests) after implementation tasks
includeReview	Boolean !	false	Generate review tasks (review) at the end of each phase
createPr	Boolean !	false	Add PR branch names to phases for automated PR creation (one PR per phase)

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 decompose-from-spec --spec string --plan string --agents string --model-family string --include-tests boolean --include-review boolean --create-pr boolean

func (m *MyModule) Example(ctx context.Context, spec string, plan string, agents string, modelFamily string, includeTests bool, includeReview bool, createPr bool) string  {
	return dag.
			Speck().
			DecomposeFromSpec(ctx, spec, plan, agents, modelFamily, includeTests, includeReview, createPr)
}

@function
async def example(spec: str, plan: str, agents: str, model_family: str, include_tests: bool, include_review: bool, create_pr: bool) -> str:
	return await (
		dag.speck()
		.decompose_from_spec(spec, plan, agents, model_family, include_tests, include_review, create_pr)
	)

@func()
async example(spec: string, plan: string, agents: string, modelFamily: string, includeTests: boolean, includeReview: boolean, createPr: boolean): Promise<string> {
	return dag
		.speck()
		.decomposeFromSpec(spec, plan, agents, modelFamily, includeTests, includeReview, createPr)
}

editFile() 🔗

Edit a file by replacing a string. The old_string must match exactly.

Return Type

Changeset !

Arguments

Name	Type	Default Value	Description
filePath	String !	-	Path to the file relative to the workspace root
oldString	String !	-	The exact string to find and replace
newString	String !	-	The replacement string
replaceAll	Boolean !	false	Replace all occurrences (default: first only)

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 edit-file --file-path string --old-string string --new-string string --replace-all boolean

func (m *MyModule) Example(filePath string, oldString string, newString string, replaceAll bool) *dagger.Changeset  {
	return dag.
			Speck().
			EditFile(filePath, oldString, newString, replaceAll)
}

@function
def example(file_path: str, old_string: str, new_string: str, replace_all: bool) -> dagger.Changeset:
	return (
		dag.speck()
		.edit_file(file_path, old_string, new_string, replace_all)
	)

@func()
example(filePath: string, oldString: string, newString: string, replaceAll: boolean): Changeset {
	return dag
		.speck()
		.editFile(filePath, oldString, newString, replaceAll)
}

glob() 🔗

Find files in the workspace matching a glob pattern.

Return Type

String !

Arguments

Name	Type	Default Value	Description
pattern	String !	-	Glob pattern (e.g. 'src/*/.py', '*/.toml')

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 glob --pattern string

func (m *MyModule) Example(ctx context.Context, pattern string) string  {
	return dag.
			Speck().
			Glob(ctx, pattern)
}

@function
async def example(pattern: str) -> str:
	return await (
		dag.speck()
		.glob(pattern)
	)

@func()
async example(pattern: string): Promise<string> {
	return dag
		.speck()
		.glob(pattern)
}

grep() 🔗

Search file contents in the workspace using grep.

Return Type

String !

Arguments

Name	Type	Default Value	Description
pattern	String !	-	Search pattern (regex supported)
paths	String !	""	Comma-separated paths to search in (optional)
fileGlob	String !	""	Glob pattern to filter files (e.g. '*.py')
insensitive	Boolean !	false	Case-insensitive search
limit	Integer !	100	Maximum number of matching lines to return

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 grep --pattern string --paths string --file-glob string --insensitive boolean --limit integer

func (m *MyModule) Example(ctx context.Context, pattern string, paths string, fileGlob string, insensitive bool, limit int) string  {
	return dag.
			Speck().
			Grep(ctx, pattern, paths, fileGlob, insensitive, limit)
}

@function
async def example(pattern: str, paths: str, file_glob: str, insensitive: bool, limit: int) -> str:
	return await (
		dag.speck()
		.grep(pattern, paths, file_glob, insensitive, limit)
	)

@func()
async example(pattern: string, paths: string, fileGlob: string, insensitive: boolean, limit: number): Promise<string> {
	return dag
		.speck()
		.grep(pattern, paths, fileGlob, insensitive, limit)
}

plan() 🔗

Generate a technical implementation plan from a specification.

Takes a feature specification and produces a technical plan with architecture decisions, project structure, dependencies, and complexity analysis. Returns the plan as markdown text.

Return Type

String !

Arguments

Name	Type	Default Value	Description
spec	String !	-	Feature specification (output of specify)
techStack	String !	""	Tech stack hints (e.g. 'Python, FastAPI, PostgreSQL')

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 plan --spec string --tech-stack string

func (m *MyModule) Example(ctx context.Context, spec string, techStack string) string  {
	return dag.
			Speck().
			Plan(ctx, spec, techStack)
}

@function
async def example(spec: str, tech_stack: str) -> str:
	return await (
		dag.speck()
		.plan(spec, tech_stack)
	)

@func()
async example(spec: string, techStack: string): Promise<string> {
	return dag
		.speck()
		.plan(spec, techStack)
}

readFile() 🔗

Read a file from the workspace with line numbers.

Return Type

String !

Arguments

Name	Type	Default Value	Description
filePath	String !	-	Path to the file relative to the workspace root
offset	Integer !	0	Line number to start reading from (1-based)
limit	Integer !	0	Maximum number of lines to read

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 read-file --file-path string --offset integer --limit integer

func (m *MyModule) Example(ctx context.Context, filePath string, offset int, limit int) string  {
	return dag.
			Speck().
			ReadFile(ctx, filePath, offset, limit)
}

@function
async def example(file_path: str, offset: int, limit: int) -> str:
	return await (
		dag.speck()
		.read_file(file_path, offset, limit)
	)

@func()
async example(filePath: string, offset: number, limit: number): Promise<string> {
	return dag
		.speck()
		.readFile(filePath, offset, limit)
}

specify() 🔗

Generate a feature specification from a prompt or GitHub issue.

Produces a structured specification with user stories, acceptance criteria, functional requirements, and success criteria. Returns the specification as markdown text.

Return Type

String !

Arguments

Name	Type	Default Value	Description
prompt	String !	""	Natural language description of what to build
issueId	Integer !	0	GitHub issue number (alternative to prompt)
repository	String !	""	GitHub repository URL (required if using issue_id)
githubToken	Secret	null	GitHub token (required if using issue_id)

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 specify --prompt string --issue-id integer --repository string

func (m *MyModule) Example(ctx context.Context, prompt string, issueId int, repository string) string  {
	return dag.
			Speck().
			Specify(ctx, prompt, issueId, repository)
}

@function
async def example(prompt: str, issue_id: int, repository: str) -> str:
	return await (
		dag.speck()
		.specify(prompt, issue_id, repository)
	)

@func()
async example(prompt: string, issueId: number, repository: string): Promise<string> {
	return dag
		.speck()
		.specify(prompt, issueId, repository)
}

task() 🔗

Launch a sub-agent for research or focused analysis.

The sub-agent has read-only access to the workspace. Use this for exploring the codebase to inform specification and planning.

Return Type

String !

Arguments

Name	Type	Default Value	Description
description	String !	-	Short description of the sub-task
prompt	String !	-	Detailed prompt for the sub-agent

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 task --description string --prompt string

func (m *MyModule) Example(ctx context.Context, description string, prompt string) string  {
	return dag.
			Speck().
			Task(ctx, description, prompt)
}

@function
async def example(description: str, prompt: str) -> str:
	return await (
		dag.speck()
		.task(description, prompt)
	)

@func()
async example(description: string, prompt: string): Promise<string> {
	return dag
		.speck()
		.task(description, prompt)
}

writeFile() 🔗

Create or overwrite a file in the workspace.

Return Type

Changeset !

Arguments

Name	Type	Default Value	Description
filePath	String !	-	Path to the file relative to the workspace root
contents	String !	-	The full file contents to write

Example

dagger -m github.com/telchak/daggerverse/speck@b7e5030da94a1004c04cde234845538e9502bcc5 call \
 write-file --file-path string --contents string

func (m *MyModule) Example(filePath string, contents string) *dagger.Changeset  {
	return dag.
			Speck().
			WriteFile(filePath, contents)
}

@function
def example(file_path: str, contents: str) -> dagger.Changeset:
	return (
		dag.speck()
		.write_file(file_path, contents)
	)

@func()
example(filePath: string, contents: string): Changeset {
	return dag
		.speck()
		.writeFile(filePath, contents)
}