Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/8BitTacoSupreme/flowstate/llms.txt

Use this file to discover all available pages before exploring further.

flowstate context reads the current flowstate.json and regenerates all five context files from your saved interview answers without re-running research, strategy, or any other pipeline step. It is the fastest way to apply manual edits to your project state — edit flowstate.json directly, then run flowstate context to propagate your changes.

​
Synopsis

flowstate context [OPTIONS]

​
Options

​
--root PATH
string
default:"."
Project root directory. Context files are written relative to this path.

​
Context files generated

flowstate context always writes all five files, creating parent directories if they do not exist:
FileConsumerContent
.planning/PROJECT.mdGSDProject name, problem statement, 10x vision, architecture pattern, test coverage target, milestones
.planning/ROADMAP.mdGSDPhase-by-phase breakdown generated from the milestones list, each with goal, deliverables, and acceptance criteria fields
.planning/config.jsonGSDWorkflow preferences: mode, granularity, auto_commit, verification
.claude/CLAUDE.mdAll toolsProject context summary: problem, vision, architecture, active tools, and a pointer to the roadmap
research/brief.mdResearch adapterStructured research questions split by topic from research_focus, with architecture context

​
When to use

  • After editing flowstate.json manually — change core_problem, milestones, architecture_pattern, or any other interview field and regenerate without a full re-run.
  • After changing project preferences — update the project name or model preferences in state, then refresh context files.
  • To reset context for a new phase — if you have updated milestones or pivoted scope, regenerate to reflect the new direction before running flowstate launch gsd.
  • To repair missing files — if a context file was accidentally deleted, flowstate context recreates it from state without any LLM calls.

​
Example output

flowstate context

5 context files written:
  .planning/PROJECT.md
  .planning/ROADMAP.md
  .planning/config.json
  .claude/CLAUDE.md
  research/brief.md

​
Example

Regenerate from the default project root: 
flowstate context
Regenerate for a project in a non-default directory: 
flowstate context --root ~/projects/my-app

​
Notes

  • Context generation is deterministic and runs in under one second. No claude CLI calls are made.
  • flowstate context is equivalent to step 1/5 of flowstate init run in isolation.
  • The research/brief.md file is consumed by the Research adapter on the next flowstate init run; regenerating it here means the next research run will reflect any changes to research_focus or architecture_pattern.

Build docs developers (and LLMs) love

Get started for freeTalk to us