weekly-recap

name	weekly-recap
description	Comprehensive weekly performance recap with portfolio review, trade analysis, and lessons learned
tags	["content","weekly","recap","performance","reporting"]
agent	writer
var	${var} can set special focus areas for the recap. If empty, produce standard weekly recap.

weekly-recap

Priority: P1 (runs weekly) Schedule: Sunday 16:00 UTC Output: Weekly recap in memory/content/recaps/

Instructions

You are executing the weekly-recap skill for the Writer agent.

Step 1: Aggregate Weekly Data

Read and compile from the entire week:

• memory/content/digests/ — all daily digests from the week
• memory/tx-log.json — all trades this week
• memory/positions.json — current positions
• memory/portfolio-state.json — portfolio state
• memory/analyst-log.json — signals and decisions
• memory/risk-log.json — risk events
• memory/research/trends/latest.json — trend context

Step 2: Performance Analysis

Calculate:

• Portfolio return: Start-of-week value vs. end-of-week
• Win rate: Profitable trades / total trades
• Best trade: Highest return trade
• Worst trade: Lowest return trade
• Sharpe ratio (weekly): Risk-adjusted return
• Max drawdown during the week
• Signal accuracy: Analyst signals that were profitable vs. not

Step 3: Write Recap

Structure (600-1000 words):

# Weekly Recap — {start_date} to {end_date}

## Performance Summary
| Metric | Value | vs. Last Week |
|--------|-------|---------------|
| Portfolio Value | ${total} | {change}% |
| Week Return | {pct}% | {delta} |
| Trades | {N} | {delta} |
| Win Rate | {pct}% | {delta} |
| Max Drawdown | {pct}% | - |

## Market Context
{2-3 sentences on overall market this week}

## Trade Review
### Best: {token} +{pct}%
{What went right, entry/exit analysis}

### Worst: {token} -{pct}%
{What went wrong, lessons}

## Signal Quality
- Received: {N} signals
- Acted on: {M} ({pct}% acceptance)
- Profitable: {K} of {M} ({win_rate}%)
- Notable misses: {signals we should have acted on}

## Lessons Learned
1. {Concrete lesson from this week's activity}
2. {Another lesson}

## Next Week Plan
- {Key levels to watch}
- {Upcoming events (unlocks, votes, launches)}
- {Strategy adjustments based on lessons}

Step 4: Save and Distribute

1. Write to memory/content/recaps/{YYYY-WW}.md
2. Post key metrics to mesh for all agents
3. Send notification summary via ./notify

Anti-Patterns

• Do NOT sugarcoat poor performance. Honest review drives improvement.
• Do NOT include hindsight trades ("we should have bought X"). Focus on process.
• Do NOT exceed 1000 words. Be analytical, not verbose.

Exit Codes

• SKILL_OK — recap generated
• SKILL_PARTIAL — some data incomplete
• SKILL_FAIL — could not generate recap

Output

Commit message format: writer: weekly-recap — week {WW} return {pct}% (win rate {win_rate}%)