impact

/st4ck:impact

Analyze a local code diff and surface every QA test whose components cite the changed lines. Queue a test_design_change dev_task per affected component, assigned to QA.

When this skill activates

Trigger	Behavior
User runs /st4ck:impact	Full skill; ask whether to invoke propose
PostToolUse hook on Edit/Write hits a cited path (opt-in hooks/post-edit-scan.sh)	Auto-trigger silently; surface dev_tasks only if queued
User asks "what tests does my diff break?"	Activate; usually skip propose (read-only)

First actions — MANDATORY

1. git rev-parse --is-inside-work-tree must return true. Else error: "This skill needs a local git checkout. Run from inside the project repo."

2. Identify diff scope. Default git diff HEAD^ HEAD. --base develop → git diff develop...HEAD. --staged → git diff --cached.

3. Capture metadata: repo_root (git rev-parse --show-toplevel), branch (git rev-parse --abbrev-ref HEAD), before_sha/after_sha, changed_files (git diff --name-only <range>), per-file changed_line_ranges (parse git diff -U0 <range> -- <path> hunk headers @@ -OLD,L +NEW,L @@).

Step-by-step

1. Build changed_files

[{ "path": "src/pages/X.tsx", "changed_line_ranges": [{"start": 142, "end": 148}] }]

2. Call get_tests_affected_by_diff

Returns { affected_tests, affected_components, suggested_runs }. If affected_components.length === 0 → exit silently with "No QA tests cite the changed lines."

3. Per affected component, create a dev_task

create_dev_task({
  title: "Test design change: <component_name>",
  description: <brief 2-line summary>,
  source_type: 'test_design_change',
  source_id: <component_id>,
  assigned_team: 'qa',
  priority: 'medium',
  body: {
    repo_root, branch, before_sha, after_sha,
    affected_citations: [...],   // §5.2 schema
    suggested_reviewer_actions: [
      "read <changed file> at HEAD — confirm component's target UI is gone",
      "if flow retired: propose archiving test",
      "if flow moved: re-record against new location"
    ]
  }
})

Server fires test_design_change_queued on test_coverage_events (plan §5.6) — QA Kanban + active authoring sessions pick it up. No manual broadcast.

4. Optional: invoke propose.md

Only if user passed --propose. See propose.md in this skill folder. Default: do NOT auto-invoke — LLM-driven re-author per component is expensive.

Safety rules

• NEVER modify test or component records yourself. Queue dev_tasks; QA reviews.
• NEVER bypass line-range overlap. Always provide changed_line_ranges — path-only matches spam the queue.
• Debounce auto-trigger. Don't recreate the same dev_task within 10s; server coalesces, you also rate-limit in hooks.
• NEVER commit on behalf of the user. Read git state, don't write.
• NEVER run the runner from this skill. Surface suggested_runs; user decides.

Output

Impact analysis (branch: <branch>, range: <before_sha>..<after_sha>):
  - <N> affected components
  - <M> affected tests
  - <K> dev_tasks queued (assigned_team='qa')
  - Suggested runs:
    npx st4ck@latest run <test_id> <base_url>

If propose ran, append Proposed diffs attached to <P> of <K> dev_tasks (confidence: low/medium/high).

Skip conditions

• No files in test_components.selector_notes.source_citations touched → MCP returns zero; exit silently.
• Mid-merge/rebase (MERGE_HEAD / REBASE_HEAD present) → git diff inconsistent; skip with note.
• Initial commit (no HEAD^) → fall back to git diff (working tree vs HEAD) or skip.