Mission Templates

Mission templates let you start a fleet run with a single command, no LLM decomposition needed. Each template is a pre-defined set of missions that works on any codebase. FleetSpark ships with 6 built-in templates.

Using a template

# List all available templates
fleet command --template list

# Start a fleet with a template
fleet command --template test-coverage

Fleet creates missions from the template, assigns them to ships, and runs them exactly like a manually planned fleet.

Built-in templates

`test-coverage`

Improve test coverage across the project. 4 missions run in parallel.

Mission	What it does
M1	Audit existing tests — find untested modules and coverage gaps
M2	Write unit tests for core utilities and helper functions
M3	Write integration tests for API routes and service boundaries
M4	Add edge-case and error-path tests for critical paths

fleet command --template test-coverage

`security-audit`

Run a comprehensive security review. 3 missions.

Mission	What it does
M1	OWASP Top 10 review — injection, XSS, auth weaknesses, insecure defaults
M2	Dependency audit — flag vulnerable packages, outdated deps, CVE matches
M3	Secret scanning — detect hardcoded credentials, API keys, and tokens

fleet command --template security-audit

`api-docs`

Generate comprehensive API documentation. 3 missions.

Mission	What it does
M1	Extract OpenAPI/Swagger spec from routes and controllers
M2	Write endpoint reference docs with parameters, responses, and errors
M3	Add usage examples and code samples for each endpoint

fleet command --template api-docs

`dependency-update`

Update project dependencies to latest versions. 3 missions.

Mission	What it does
M1	Update dependencies and resolve breaking API changes
M2	Fix type errors and compatibility issues introduced by updates
M3	Verify CI passes and update lock files

fleet command --template dependency-update

`refactor`

Improve codebase structure and maintainability. 4 missions.

Mission	What it does
M1	Extract shared logic into reusable utilities
M2	Simplify complex functions — reduce cyclomatic complexity
M3	Remove dead code, unused exports, and stale comments
M4	Add or improve TypeScript types for public interfaces

fleet command --template refactor

`drsti-dev-flow`

Governed development workflow: spec, implementation, and peer-review — one mission per phase. Designed for teams using the drsti-dev-flow maturity model but works on any project.

Mission	What it does
M1 (spec)	Read the adapter config and codebase context. Run a pre-proposal coordination check. Write a spec covering scope, contracts, acceptance criteria, and risks. Self-review and commit.
M2 (impl)	Read the M1 spec. Implement within declared scope, write or update tests, self-review against the spec.
M3 (review)	Peer-review the M2 implementation against the M1 spec — spec compliance, test coverage, scope drift. Record findings and approve only when all blockers are resolved.

Each brief instructs the agent to rename its branch before starting (e.g. feature/my-feature-spec). Run the template once per feature.

fleet command --template drsti-dev-flow

Note: M2 depends on M1, M3 depends on M2 — missions run sequentially, not in parallel. This is intentional: each phase gates the next.

Combining templates with configuration

Templates use the agents and settings already defined in .fleet/config.yml. You can adjust the number of ships, agent type, and other settings before running a template:

ship:
  adapter: claude-code   # all template missions use this agent
commander:
  maxConcurrentMissions: 3

What’s coming

The template system is designed to be extensible. In a future release you will be able to:

Load templates from local YAML files (--template-file my-template.yml)
Install community templates as npm packages
Publish your own templates to share with your team

For now, all templates are built into the fleetspark CLI and work offline.