AGENT SKILL

deployment-and-operations

Install, start, containerize, operate, and safely load-test the APOD Flask service.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
deployment-and-operations
license
Apache 2.0
metadata
{ "disco-role": "operating" }
description
Install, start, containerize, operate, and safely load-test the APOD Flask service.
disable-model-invocation
true

02 资源文件 6

03 SKILL.md 正文

Deployment and Operations

Use this sub-skill when the task is to install the public runtime dependencies, start APOD locally, run the production Gunicorn server, build or compose the container, perform a bounded startup check, or run the optional weighted Locust profile. The service is Python >=3.12 and the package metadata version is 1.1.0.

Route here for

  • uv sync, public dependency installation, or a failed pip install -e .
  • gunicorn application:app and port 5000 startup
  • Docker image builds, Compose port mappings, and container startup
  • Static/template asset checks and bounded service smoke checks
  • A deliberately short Locust run against an explicitly named target

Route away

Read next

  1. Read installation for Python >=3.12, uv, public pip alternatives, and the flat-layout packaging caveat.
  2. Read container deployment for the source Docker behavior and the adapted Compose template.
  3. Read operations for bounded startup checks, Gunicorn, shutdown, and the optional Locust profile.
  4. Read troubleshooting for packaging, import, port, asset, Docker, network, and credential-bound failures.

Safety boundaries

  • External NASA APOD access is required for live /v1/apod scraping. Import, /, and static-asset checks are the default bounded checks; do not start a live load test as part of ordinary setup.
  • Locust is a development-only optional group, not a minimum runtime dependency. Every load run must supply an explicit host and finite users, spawn rate, and run time.
  • alchemy_api.key is credential-bound. Without that file, concept tagging is disabled by design; never bake credentials into an image, Compose asset, or command transcript.
  • The application must run with its working directory at the checkout root so templates/, static/, and the optional key lookup resolve as intended.

Canonical production command

From a real checkout after dependencies are available, the production entry point is exactly:

gunicorn -w 4 -b 0.0.0.0:5000 --access-logfile - application:app

Use uv run gunicorn ... when uv owns the environment. The command starts a long-running service; use the bounded checks in operations first when the request is only to validate installation.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

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).

★ 351