From logvalet
Generate a weekly or daily digest of Backlog project activity: completed issues, newly started work, blocked items, and active issue counts — deterministic, no LLM needed. TRIGGER when: user says "週次ダイジェスト", "日次ダイジェスト", "今週の進捗", "weekly digest", "daily digest", "今週のまとめ", "進捗まとめ", "先週の振り返り", "日報ネタ", "週報ネタ", "今日の進捗", "プロジェクトの今週の動き", "weekly summary", "daily summary", "weekly progress", "今週完了したもの", "今日のアクティビティ", "定期レポート", "periodic digest", "進捗サマリー". DO NOT TRIGGER when: user wants a full activity report with user/team breakdown (use report) or wants anomaly detection (use intelligence). Complements: Combine with /logvalet:report for a comprehensive periodic review.
How this skill is triggered — by the user, by Claude, or both
Slash command
/logvalet:digest-periodicThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
`lv digest weekly` / `lv digest daily` を材料に、LLM が人間向けのサマリーとハイライトを生成する。
lv digest weekly / lv digest daily を材料に、LLM が人間向けのサマリーとハイライトを生成する。
For full logvalet CLI documentation, see the
logvaletskill.
Use logvalet-digest-periodic when you need to:
If the user provides a project key and period (weekly/daily), use them directly.
If not provided, ask:
lv project list -f md)weekly or dailydaily: specific date (default: today)For weekly digest:
lv digest weekly -k PROJECT_KEY -f json
For daily digest:
lv digest daily -k PROJECT_KEY -f json
lv digest daily -k PROJECT_KEY --since YYYY-MM-DD --until YYYY-MM-DD -f json
Note: Use --since / --until flags (not --date) when specifying a custom date range.
The digest JSON includes structured fields to guide LLM reasoning:
summary: pre-computed counts and metrics (completed, new, in-progress, overdue, stale)completed_issues: list of issues closed in the periodnew_issues: list of issues created in the periodstatus_changes: issue status transitionsactive_comments: comment activityllm_hints: LLM-facing guidance on what to highlight, what risks to surface, and what actions to suggestAlways read and apply llm_hints — they encode deterministic signals from the CLI (e.g., "N issues are overdue", "M issues are stale") that should directly inform the narrative. The llm_hints field removes guesswork and ensures the summary accurately reflects project state.
Using the digest data and llm_hints, produce a Markdown summary:
## 週次ダイジェスト — PROJECT_KEY (YYYY/MM/DD - YYYY/MM/DD)
### ハイライト
- 完了: N件(PROJ-XXX, PROJ-YYY)
- 新規開始: N件
- 注目: PROJ-ZZZ が期限に近づいています
### リスク・懸念
- N件が停滞中(7日以上更新なし)
- 期限超過: N件
### 次のアクション
- PROJ-ZZZ のステータス確認
- UserName の過負荷解消のためのタスク再配分検討
---
詳細を確認しますか? 特定の課題に絞りますか?
For daily digest, adjust the header to:
## 日次ダイジェスト — PROJECT_KEY (YYYY/MM/DD)
After presenting the summary, offer to:
llm_hints in the digest output is the primary signal for risk and action items — always surface content flagged theresummary fields provide pre-computed counts — use them directly rather than re-counting from raw arrayslv digest weekly -k PROJ1 -k PROJ2 -f json with multiple -k flags--date flag — use --since / --until for custom date rangesllm_hints — they contain pre-analyzed signals that improve summary accuracysummary already provides themnpx claudepluginhub youyo/claude-plugins --plugin logvaletActivate for: weekly status, project status, milestone, blocker, RAG status, executive dashboard, portfolio view, progress report, what's blocking, how are we doing on, what's the status of, where are we with, is this on track, milestone check, what have we completed, what's in progress. NOT for: daily digest (use digest), brain dump prioritisation (use task-intelligence).
Generates project status reports from Jira issues using JQL, categorizes by status/priority, analyzes blockers/progress, formats for executives/teams, and publishes to Confluence.
Generates daily standup reports aggregating 24h GitHub activity (PRs, commits, reviews) and Linear issues by status (In Progress, Todo, blockers). Ideal for morning prep or activity summaries.