sf-tasks

SF Tasks — Task Backlog from an Audit

Converts an audit (audit.json) into a ready-to-use backlog: tasks.json (machine-readable) and tasks.md (a checklist organized by priority). The pipeline is configurable: what to include, how to group items, and which priorities and effort estimates to assign.

When to Use It

• "Create tasks from this export / audit";
• "Build a backlog for developers" or "What should be fixed first?";
• "I need tasks.md / a prioritized task list."

Workflow

1. Obtain the audit. If audit.json is available, go directly to step 3. If only SF exports are available, run the audit first:

   seohead sf run --exports-dir ./exports --out ./report
   

2. (Optional) Configure the pipeline in config.json → tasks_pipeline:
   • include_severities — which levels to include (all by default);
   • group_by — check (one task per issue type) or issue (by URL);
   • priority_map — severity → P1/P2/P3; effort_map — severity → effort;
   • max_urls_per_task, min_occurrences, include_checks/exclude_checks.
3. Build the tasks:

   seohead sf tasks --json ./report/audit.json --out ./report --config config.json
   # Or generate them together with the audit in a single command:
   seohead sf run --exports-dir ./exports --out ./report --tasks
   

   Through MCP, use the sf_audit_tasks {json_path, out, config?} tool.
4. Deliver the result. tasks.md is a checklist organized by P1/P2/P3 with a "how to fix" field; for broken links, it includes a destination ← source · position · XPath list. Discuss P1 first, estimate the total scope (summary.by_priority), and attach tasks.json for the tracker.

Task Format (tasks.json)

id, check, priority (P1/P2/P3), severity, effort, title, fix_hint, affected_count, occurrences, urls[] (subject to the limit), and, for links, broken_links[] with location details. Groups and priorities are deterministic, so the backlog can be diffed between runs.