cli-local-development
Use when a coding agent needs to run local Dagster OSS CLI commands for project scaffolding, workspace/code-location loading, local development, definitions validation, asset materialization, job execution/launch/listing, schedule/sensor operations, instance inspection, or debug export/import.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- cli-local-development
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Use when a coding agent needs to run local Dagster OSS CLI commands for project scaffolding, workspace/code-location loading, local development, definitions validation, asset materialization, job execution/launch/listing, schedule/sensor operations, instance inspection, or debug export/import.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/troubleshooting.md8.9 KB
- 参考文档references/workflows.md12.1 KB
- 脚本scripts/dagster_cli_doctor.py5.2 KB
03 SKILL.md 正文
CLI Local Development
Use this sub-skill for local dagster command-line workflows in an installed Dagster Python environment. It covers commands that inspect or run user code locally, validate code locations, start dagster dev, execute jobs/assets, manage schedule/sensor state, inspect the current instance, and collect debug artifacts.
Route First
- Use this sub-skill for
dagster project,dagster dev,dagster definitions validate,dagster asset,dagster job,dagster schedule,dagster sensor,dagster instance, anddagster debugusage. - Route webserver startup internals, GraphQL API calls, GraphQL client code, and webserver debug viewing to
../graphql-and-webserver/SKILL.mdif that sub-skill exists. - Route production instance storage, daemon deployment, run launcher configuration, and operational deployment configuration to
../deployment-operations/SKILL.mdif that sub-skill exists. - Route
dgcomponent/project tooling to../components-projects/SKILL.mdif that sub-skill exists; many legacydagstercommands now print supersession warnings but remain useful for local CLI support.
Start Here
- Confirm the environment can import Dagster and expose the needed commands with
python scripts/dagster_cli_doctor.pyfrom this sub-skill directory, or run its path from elsewhere. - Choose a loading target deliberately:
-f/--python-file,-m/--module-name,--package-name,-a/--attribute,-d/--working-directory, or-w/--workspace. - Validate loading before executing work:
dagster definitions validate ...with the same target flags. - For run config, use either repeated
-c/--configYAML files/patterns or a single--config-jsonstring, never both. - For persistent schedule/sensor/instance behavior, set
DAGSTER_HOME; otherwise many commands use an ephemeral instance and state may disappear after the command exits.
References
- Local CLI workflows for concrete command patterns, target selection, config merging, project scaffolding, local dev, validation, assets, jobs, schedules, sensors, instances, and debug commands.
- Troubleshooting for import failures, target/load errors, optional dependency gaps, config misuse, partitions, schedule/sensor state, and instance surprises.
- Dagster CLI doctor script for a safe environment and
--helpsmoke check that does not start services or execute user code.
Safety Notes
dagster devstarts a local webserver and daemon; do not run it as a smoke test unless the user asked to start services.dagster asset wipe, schedule/sensor start/stop,dagster instance migrate,dagster instance reindex, and debug import mutate local Dagster state; confirm intent before running them.dagster project from-examplemay download examples; preferdagster project list-examplesor scaffold commands for offline checks.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).