zigflow graph inject

Inject a workflow graph into a file between marker comments

Synopsis

Inject a rendered workflow graph into one or more target files.

For each target file the command finds a pair of marker comments, replaces everything between them with a freshly generated graph, then writes the result back. Running inject is idempotent: if the graph is already current the file content does not change.

Auto-detect form (no --workflow flag):

zigflow graph inject <target-file> [target-file...]

The workflow file path is read from the start marker embedded in each target file. Embed a path like this:

<!-- ZIGFLOW_GRAPH_START ./workflow.yaml -->
<!-- ZIGFLOW_GRAPH_END -->

When no path is embedded the default "workflow.yaml" is used, resolved relative to the target file's directory. If a target file contains no ZIGFLOW_GRAPH_START marker it is skipped silently, so the command is safe to use with pass_filenames: true in pre-commit hooks.

Explicit form (with --workflow flag):

zigflow graph inject --workflow <workflow-file> <target-file> [target-file...]

All targets use the supplied workflow file and the --start-marker value.

Default markers:

<!-- ZIGFLOW_GRAPH_START -->
<!-- ZIGFLOW_GRAPH_END -->

zigflow graph inject [flags] <target-file> [target-file...]

Options

  -e, --end-marker string     End marker comment (default "<!-- ZIGFLOW_GRAPH_END -->")
  -h, --help                  help for inject
  -o, --output string         Output format (mermaid) (default "mermaid")
  -s, --start-marker string   Start marker comment (requires --workflow) (default "<!-- ZIGFLOW_GRAPH_START -->")
  -w, --workflow string       Explicit workflow file (auto-detected from start marker if not set)

Options inherited from parent commands

  -l, --log-level string   Set log level (default "info")

zigflow graph inject