explain-why

name	Explain Why
description	Answer 'why is X in my priorities/activities/signals/drift alerts?' by tracing a brief item back to its source signals and breadcrumbs.

When to Use

Invoke this skill when the user asks a question matching any of these patterns (case-insensitive):

• "why is [item] in my priorities?"
• "why did [item] show up?"
• "where does [item] come from?"
• "what signals produced [item]?"
• "why am I seeing [item]?"
• Any question containing "why" paired with a named item from a brief (priority, activity, drift alert, or signal).

Output

Render to terminal only. Do NOT send email or write to memory.

📍 Why "[item name]"?

Source breadcrumb: [source_breadcrumb]
Urgency / relevance: [urgency or relevance score]
Pass reason: [pass_reason, if applicable]

This item was produced by: [skill name]
Found in: [memory file and field path]

If multiple breadcrumbs exist (e.g., item appears in both priorities and activities), show all of them.

Instructions

Step 1 — Parse the item name

Extract the subject noun phrase from the user's question. For example:

• "Why is auth migration in my priorities?" → auth migration
• "Why did the cost review show up?" → cost review
• "Where does ARM CI pipeline come from?" → ARM CI pipeline

Step 2 — Search memory (most-recent first)

Search the following memory files in order, stopping as soon as the item is found. Use the dayarc-memory skill to read each file. Match by fuzzy keyword overlap (not exact string match).

1. daily/daily-profile-{today}.json — check these fields:

   • priorities_today[].description + priorities_today[].source_breadcrumb (produced by infer_priorities)
   • unfinished[].description + unfinished[].source_breadcrumb (produced by learn_user_profile)
   • active_threads[].subject + active_threads[].source_breadcrumb

2. weekly-summary-current.json — check:

   • themes[].label + any breadcrumbs stored under activities for that theme
   • stuck[].description + stuck[].source_breadcrumb

3. monthly-summary.json — check:

   • themes[].label + activity breadcrumbs
   • stuck[].description + stuck[].source_breadcrumb

4. Recent daily profiles — if today's profile has no match, try the previous 3 days: daily/daily-profile-{yesterday}.json, {day-2}.json, {day-3}.json. In each, check priorities_today, unfinished, and active_threads as above.

Step 3 — If item not found in memory

If no match is found across all memory files, say:

I couldn't find "[item]" in your recent memory (last 3 days, current weekly, current monthly). It may be from an older brief or may have been removed. Try asking about a specific date, e.g. "why was auth migration in my priorities on March 20?"

Step 4 — Render breadcrumbs

Once the item is found, render all available source fields:

• For priorities (infer_priorities output): show source_breadcrumb and urgency.
• For activities (classify_activity output): show source_breadcrumb per activity in the theme cluster.
• For drift alerts (detect_drift output): show source_breadcrumb (the priority that was neglected) and days_inactive.
• For signals (filter_signals output): show source_breadcrumb, relevance, and pass_reason.

Apply the breadcrumb quality rule: if a breadcrumb contains a Teams meeting thread link (19:meeting_...@thread.v2), flag it with ⚠️ and include whatever fallback context (meeting title, date, participants) is stored alongside it.

Step 5 — Cite the source skill

End the response with a one-line attribution:

Produced by [skill-name] · found in [memory-file] › [field-path]

Example:

Produced by infer_priorities · found in daily-profile-2026-03-28.json › priorities_today[1]