메뉴
Run note quality maintenance — invoke /inspect to diagnose vault metadata, then /repair to fix queued issues. Use when the user wants weekly or periodic Rill vault health maintenance.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Generate today's Daily Note (or one for a specified date) at reports/daily/YYYY-MM-DD.md by aggregating the day's situation across journals, tasks, workspaces, knowledge, activity-log, and any plugin briefing hooks. Fully automated, no interaction. Use when the user asks for a daily briefing, situational analysis, or "what's going on today/yesterday".
Ingest a single Twitter/X tweet URL into inbox/tweets/, fetch its content via fetch-tweet.sh, and generate an organized version with frontmatter and Markdown body in _organized/. Combines `rill clip` with the tweet branch of /distill Phase 2 in one shot. Use when the user supplies a single tweet URL to capture.
Quantitatively evaluate the search quality and efficiency of the Rill knowledge base via natural exploration vs Deep Path ground truth (ADR-050). Run monthly. Computes precision, recall, metadata_contribution, token cost. Use when the user wants to verify whether note structure actually works in practice.
Start (or resume) a Rill workspace from a theme, journal file, or workspace id. Automatically collects related vault context and creates a workspace/{id}/ directory for an interactive Deep Think session. Use when the user wants to begin or continue a workspace-scoped exploration.
Diagnose Rill PKM taxonomy health, metadata accuracy, and file integrity. Run weekly. Detects issues and appends mismatched files to knowledge/.refresh-queue for repair by /repair.
Run the daily morning routine — invoke /briefing (Daily Note) followed by /newsletter (research report) inline within the current session. Use when the user starts the day or explicitly asks for the morning routine.
| name | maintain |
| description | Run note quality maintenance — invoke /inspect to diagnose vault metadata, then /repair to fix queued issues. Use when the user wants weekly or periodic Rill vault health maintenance. |
Conduct all conversation with the user — and write all generated output — in the language defined by .claude/rules/personal-language.md (or the user's input language if that file is absent). Follow the language rules in full — exceptions and translation quality are defined in the Language Rules of .claude/rules/rill-core.md and the vault's personal-*.md overrides, never restated per skill. The English instructions below are for skill clarity, not for output style.
Tool references in this skill (
sub-agent, "skill invocation") describe intent, not Claude-specific tool calls. Each harness should map them to its native equivalent — Claude Code uses its built-in Skill / Agent tools as named; Codex CLI uses its own skill invocation mechanism (mention / implicit //skills) and sub-agent mechanism as appropriate.
Runs note quality inspection and repair sequentially. Diagnoses issues with /inspect, then fixes them with /repair if any are found.
$ARGUMENTS — None
Invoke the inspect skill (via the harness's skill mechanism).
/inspect diagnoses taxonomy health + metadata accuracy + consistency, and adds problematic files to knowledge/.refresh-queue.
Read knowledge/.refresh-queue and check the count.
Invoke the repair skill (via the harness's skill mechanism).
/repair reads files from .refresh-queue and batch-normalizes tags/mentions/type.
/inspect and /repair use sub-agents internally, but are not as heavy as /distill. The combined context consumption of the 2 skills is manageable (~50-80K tokens). To avoid the process startup overhead of spawning a new harness session, they are run sequentially inline within the current session.
