Weekly Review — Weekly Persona Review

When this skill is activated

1. Automatic: HEARTBEAT executes this skill every Monday at 9am.
2. Manual: user sends /weekly-review or requests a weekly review.

Complete Flow

HEARTBEAT (Monday, 9am) or manual trigger
        │
        ▼
Load list of active personas from USER.md
        │
        ▼
For each active persona:
  life_log.get_persona_summary(persona, days=7)
        │
        ▼
Calculate usage metrics per persona
        │
        ▼
Compare usage pattern with current_weight of each persona
        │
        ▼
Generate report with adjustment suggestions
        │
        ├── No suggestions → inform weights are adequate
        │
        └── Has suggestions → present report to user
                │
                ▼
          Wait for user confirmation
                │
                ├── User confirms all → apply all adjustments via weight-system
                ├── User confirms partially → apply only confirmed ones
                └── User refuses → don't apply any adjustment

Step 1 — Collect data from life_log

For each active persona listed in USER.md, call:

life_log.get_persona_summary(persona="{slug}", days=7)

Summary returns:

• total_interactions: total number of interactions in last 7 days
• intents_used: list of executed intents with count (e.g., {"code_review": 12, "planning": 3})
• tools_used: list of called tools with count (e.g., {"github": 8, "calendar": 2})
• active_hours: list of hours of day with most activity (e.g., [9, 10, 14, 15])
• last_interaction: timestamp of last interaction

If a persona has no interactions in last 7 days, record total_interactions=0.

Step 2 — Calculate usage metrics

For each persona, calculate:

Step 3 — Compare with current weights

For each persona, compare usage rate with current_weight:

Weight increase suggestion criteria

Suggest increase of current_weight if:

• usage_rate > current_weight + 0.15 (persona being used much more than weight suggests)
• total_interactions >= 20 in the week (consistent and expressive usage)

Suggested value: min(current_weight + 0.1, 1.0)

Weight reduction suggestion criteria

Suggest reduction of current_weight if:

• usage_rate < current_weight - 0.2 (persona being used much less than weight suggests)
• total_interactions <= 3 in the week (very low usage)
• current_weight > 0.3 (don't reduce personas already with low weight)

Suggested value: max(current_weight - 0.1, 0.0)

No suggestion

Keep current weight if none of the above criteria are met.

Step 4 — Generate report

Assemble report in clear and accessible language for user:

{locale:title}

{locale:separator}

{For each persona with interactions:}

{locale:persona_header}
   {locale:interactions}
   {locale:most_used_for}
   {locale:favorite_tool}
   {locale:peak_hours}
   {locale:current_weight}
   {If has suggestion:}
   {locale:suggestion}

{locale:separator}

{If has personas without interactions:}
{locale:inactive_title}
   {locale:inactive_item}
   {locale:inactive_item}

{locale:separator}

{If has adjustment suggestions:}
{locale:suggestions_prompt}

{If no suggestions:}
{locale:no_suggestions}

Report language rules:

• Use simple language, no technical jargon
• Replace "current_weight" with "weight" or "priority"
• Replace "intent" with "task type" or "what was done"
• Replace "tool" with "tool"
• Dates in DD/MM/YYYY format

Step 5 — Wait for confirmation before applying

Never apply adjustments without explicit user confirmation.

User response options

Individual confirmation flow (when user responds "choose")

For each suggestion, ask:

"{locale:confirm_individual}"

Wait for response before moving to next.

Step 6 — Apply confirmed adjustments via weight-system

For each confirmed adjustment, call:

weight_system.adjust_weight(
    persona_slug="{slug}",
    delta={new_weight - current_weight},
    justification="Weekly review: usage rate {usage_rate:.0%} vs weight {current_weight} — {reason}"
)

weight-system persists new current_weight in USER.md and records change with timestamp and justification in personas/{slug}/weight-history.md.

After applying all confirmed adjustments, confirm to user:

{locale:applied}

Edge Cases

Dependencies

• skills/life-log — get_persona_summary(persona, days=7) to collect usage data
• skills/weight-system — adjust_weight(persona_slug, delta, justification) to apply adjustments
• USER.md — list of active personas and their current_weight
• personas/{slug}/weight-history.md — adjustment history (written by weight-system)