-
Confirm
README.md,LICENSE, and development docs are present. -
Run validation:
go test ./packages/cli/... go run ./packages/cli/cmd/flowtrace validate --root . .flowtrace/cli-and-nvim-paths.flow.json
-
Confirm the GitHub repository is public at
github.com/ebrakke/flowtrace. -
Push the repository and create an initial tag, for example
v0.1.0. -
Verify install from outside the checkout:
go install github.com/ebrakke/flowtrace/packages/cli/cmd/flowtrace@v0.1.0
-
Keep the canonical skill at
skills/flowtrace/SKILL.md. -
Validate with
skills-ref validate ./skills/flowtracewhen available. -
After the GitHub repository is public, verify:
npx skills add https://github.com/ebrakke/flowtrace --skill flowtrace
-
Submit or wait for skills.sh indexing according to the current skills.sh process.
flowtrace buildis a deterministic search scaffold and may mark nodes withresolution: search.- High-quality flows are expected to be authored by the calling coding agent, then validated with the CLI.
- The Neovim plugin lives in
packages/nvim, so lazy.nvim installs from the public repo with a small runtimepath config block.