data-audit
Report on data completeness for the SPARTA QRA pipeline (Control -> URL -> Knowledge -> QRA). Queries DuckDB to show coverage percentages at each stage.
ソース情報
- リポジトリ
- grahama1970/agent-stack-public
- ソースの最終更新活動
- 2026年9月24日 15:51
- 検出された SKILL.md の言語
- 英語
- スター
- 0
- フォーク
- 0
インストール方法
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
ソースファイルを確認
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
ファイルエクスプローラー
4 ファイルSKILL.md を表示中
SKILL.md
ソースの指示 · 読み取り専用プレビュー- name
- data-audit
- description
- Report on data completeness for the SPARTA QRA pipeline (Control -> URL -> Knowledge -> QRA). Queries DuckDB to show coverage percentages at each stage.
- triggers
- ["audit sparta","check coverage","data completeness","pipeline status"]
- metadata
- {"short-description":"SPARTA data completeness auditor.","project-path":"${HOME}/workspace/experiments/pi-mono/.agent/skills/data-audit"}
- provides
- ["data-audit"]
- composes
- ["create-figure","task-monitor","agentic-evals"]
- disciplines
- ["data-engineering","evaluation-quality"]
# Data Audit
Audits the SPARTA pipeline data coverage.
## Usage
```bash
# Run full audit
.agent/skills/data-audit/run.sh
# Run for specific run ID (points to specific DB path if needed)
.agent/skills/data-audit/run.sh --run-id <run-id>
```
## Logic
Connects to SPARTA DuckDB and calculates coverage for:
1. **Controls**: Base set.
2. **URLs**: Controls mapped to at least one URL.
3. **Knowledge**: Controls mapped to URLs that have extracted chunks.
4. **QRA**: Controls that have generated QRA pairs.
## Visualization
After generating coverage data, offer to visualize via `/create-figure`:
```bash
# Pipeline coverage as stacked bar chart
create-figure metrics --input audit.json --output coverage.png --type bar --title "SPARTA Pipeline Coverage"
# Coverage heatmap by category
create-figure heatmap --input audit.json --output coverage-heatmap.png
```
**When to offer:** After presenting coverage percentages, ask: "Want me to chart the pipeline coverage?"
## Common Mistakes
### WRONG: Running audit without a DuckDB path when needed
```bash
./run.sh # may fail if DuckDB path is not configured
```
### RIGHT: Specify run ID or verify default DB path
```bash
./run.sh --run-id <run-id> # explicit run
```
### WRONG: Reporting raw counts without coverage percentages
```
Controls: 800, URLs: 500, QRAs: 200 # raw numbers are meaningless
```
### RIGHT: Report coverage ratios at each pipeline stage
```
Controls: 800 → URLs: 500 (62.5%) → Knowledge: 350 (43.8%) → QRA: 200 (25.0%)
```
### WRONG: Not visualizing results for stakeholder communication
```bash
./run.sh # text output only, hard to share
```
### RIGHT: Generate charts via /create-figure after audit
```bash
./run.sh > audit.json
create-figure metrics --input audit.json --output coverage.png --type bar
```
GitHubで見る