菜单
Updates an existing Airflow DAG to reflect changes in the LLD or config. Applies incremental edits, bumps version comment, and re-validates. Use when the user asks to: - Update, modify, or patch an existing DAG - Reflect an LLD change in the pipeline - Add or remove tasks from an existing DAG
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Updates an existing Sprint Backlog with new information. Reads the existing backlog, epics, and stories, then merges updated LLD requirements, changed upstream artifacts, revised team capacity, or sprint re-planning. Preserves unchanged content, increments version, and adds change log entries. Use when the user asks to update, revise, or modify existing stories or epics.
Marks Scrum stories Done and rolls epic/backlog status up once every child is complete. BLOCKS completion if any story AC checkbox is unchecked, any dependency is not Done, any referenced deliverable is missing, or any epic-level AC is still open. Uses Edit (not Write) to flip only the Status cell and AC checkboxes, preserving every other line. If the gate blocks on ANY target, NOTHING is edited — the run is atomic from the user's view. Use when the user asks to: - Mark STORY-NN-NNN complete / close a story - Close out EPIC-NN / roll up epic status - "Are we done with sprint 3 yet?" (gate run without edits is acceptable)
Generates an Airflow DAG from the LLD artifact and pipeline configs. Reads the approved LLD from outputs/, applies DAG templates, and writes a production-ready DAG file to airflow/dags/. Also known as: dag generation, pipeline scaffolding, airflow pipeline creation. Input formats: LLD markdown, DAG config YAML. Output format: Python DAG file (.py). Use when the user asks to: - Create, generate, or scaffold an Airflow DAG - Translate an LLD into a runnable pipeline - Build a DAG for bronze/silver/gold layers
Generates layer-scoped local deploy smoke artifacts: shell scripts under _infra/cd/ that re-apply DDL migrations and re-sync the Airflow DAG bag, plus a pytest smoke module under tests/integration/ that drives those scripts and re-triggers the layer DAG end-to-end. Pairs with the scrum-master deploy-validation story type 1:1. Use when the user asks to: - Author the local deploy smoke for a layer (bronze / silver / gold) - Wire up _infra/cd/ scripts (DAG sync, DDL apply) for local - Implement a deploy-validation story (STORY-NN-NNN typed deploy-validation)
Generates the Gold layer of the Patient 360 Medallion pipeline from the approved LLD §2.3 / §5.3, DMS §4, STM Silver-to-Gold sheet, DQS §2-3 (Gold rules), and a given story. Emits PySpark builder modules for the Gold consumer tables declared in LLD §5.3 (currently `patient_summary`, `patient_clinical_history`, `patient_billing_summary` — derived at runtime, not hardcoded), per-table contracts, per-table Spark Expectations DQ rule files, unit tests, and the DAG wiring for the `gold_build` task group. Modes: - Story mode: STORY-NN-NNN — emit only Gold deliverables in scope - Full mode: process every un-Done story classified as `gold`, topo-sorted by Depends On, until the Gold layer is complete Project-agnostic: the table list, source-table joins, business-invariant DQ rules, and module names are ALL read from the LLD §5.3, DMS §4, STM Silver-to-Gold sheet, and DQS §2-3 at runtime. Use when the user asks to: - Generate or update Gold layer code for a story or epic - Build the Gold consumer tables (patient_sum
Generates the Bronze config-driven ingestion layer from the approved LLD, DMS, STM, DQS, and a given story. Modes: - Story mode: read a STORY-NN-NNN, extract backtick-quoted deliverables from its Acceptance Criteria, generate only those that fall under this skill's domain of ownership. ROUTE-OUT for others. - Full mode: process every un-Done story classified as `ingestion`, topo-sorted by Depends On. Project-agnostic: the table list, critical-table designation, per-table DQ rule filenames, and runner module names are ALL read from the current run's LLD / DMS / STM / DQS / stories at runtime. No project identifiers or story IDs are hardcoded. Use when the user asks to: - Generate or update Bronze ingestion code for a story - Build the ingestion runner / factory / SparkSubmit wrapper - Generate per-table ingestion configs
| name | update-dag |
| description | Updates an existing Airflow DAG to reflect changes in the LLD or config. Applies incremental edits, bumps version comment, and re-validates. Use when the user asks to: - Update, modify, or patch an existing DAG - Reflect an LLD change in the pipeline - Add or remove tasks from an existing DAG |
| argument-hint | [dag-file-path] |
| allowed-tools | Read, Write, Edit, Grep, Glob, Bash, AskUserQuestion, Skill |
| context | fork |
You are a senior Data Engineer specialising in Apache Airflow. Your job is to apply incremental changes to an existing DAG based on an updated LLD or user instruction, without breaking existing task dependencies.
Before any file operation, run the discovery helper and substitute the returned tokens into every path this skill reads, writes, or edits:
python3 ${CLAUDE_PLUGIN_ROOT}/skills/validate-stories/scripts/status_rollup.py --mode discover
The JSON output supplies {workspace_root}, {project_root},
{project_name}, {stories_dir}, and {learnings_queue}. The plugin is
project-agnostic — never hardcode project or chapter names in edits.
Before editing a DAG, load the latest coding-patterns handbook:
PATTERNS_DIR=$(ls -d "{workspace_root}/inputs/code/v"* 2>/dev/null | sort -V | tail -1)
if [ -z "$PATTERNS_DIR" ] || [ ! -d "$PATTERNS_DIR" ]; then
echo "CRITICAL: inputs/code/v*/ not found. Run /developer-plugin:refresh-libraries to initialize the library cache."
exit 1
fi
LIBRARIES_FILE="$PATTERNS_DIR/LIBRARIES.md"
Required pattern docs for this skill:
$PATTERNS_DIR/airflow-dag-pattern.md — factory, TaskGroup, SparkSubmitOperator defaults$PATTERNS_DIR/naming-conventions.md — DAG file name, task IDs$PATTERNS_DIR/LIBRARIES.md — pinned Airflow + provider versionsLAST_VERIFIED=$(grep '^last_verified:' "$LIBRARIES_FILE" | awk '{print $2}')
TODAY=$(date -u +%Y-%m-%d)
AGE_DAYS=$(python3 -c "from datetime import date; print((date.fromisoformat('$TODAY') - date.fromisoformat('$LAST_VERIFIED')).days")
If AGE_DAYS > 30, pause and call AskUserQuestion with options Refresh now / Proceed with cached versions / Cancel. On Refresh, invoke /developer-plugin:refresh-libraries then resume.
Emit a ### References section citing consumed pattern docs + LIBRARIES.md vintage. Add a stale-cache warning if the user proceeded with cached versions.
The Skill-tool argument frequently fails to reach forked subagents. Resolve
the target via the shared resolver, which checks four sources in order:
$SKILL_ARG → {workspace_root}/.skill-arg → conversational arg → auto-mode.
# Step 1: capture the user's conversational input. Substitute the
# bracketed text below with the EXACT message the user supplied after
# the skill name; if no message was supplied, leave it as an empty
# string. This is the ONLY substitution this skill requires.
CONV_ARG='<<EXACT_CONVERSATIONAL_TEXT_FROM_USER_OR_EMPTY_STRING>>'
# Step 2: run the shared resolver. It auto-discovers the workspace
# from $PWD, so no {workspace_root} substitution is required. Output is
# two lines on stdout: the resolved value, then the source token.
read -r RESOLVED_ARG RESOLVED_SOURCE < <(
bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve_skill_arg.sh" "$CONV_ARG" \
| paste -sd' ' -
)
Print this banner as the first line of skill output:
RESOLVED TARGET: <value> (source: <SKILL_ARG | .skill-arg | conversational | __AUTO__>)
If $RESOLVED_SOURCE == EMPTY, fall through to the skill's existing
clarification step (typically AskUserQuestion). DO NOT ask the user before
running this resolver.
{workspace_root}/outputs/lld/v*/ to understand what changedUse AskUserQuestion to confirm scope if the diff is ambiguous.
Invoke /developer-plugin:validate-dag on the updated file.