AGENT SKILL

repo-maintenance

Maintain the Langflow monorepo safely across tests, formatting, generated files, versioning, migrations, releases, docs, and CI-adjacent checks.

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

01 前言 Front Matter

name
repo-maintenance
license
MIT
metadata
{ "disco-role": "operating" }
description
Maintain the Langflow monorepo safely across tests, formatting, generated files, versioning, migrations, releases, docs, and CI-adjacent checks.
disable-model-invocation
true

02 资源文件 4

03 SKILL.md 正文

Langflow Repo Maintenance

Use this sub-skill for repository-wide Langflow maintenance: setup, focused validation, formatting, linting, package/workspace version updates, generated artifacts, Alembic and extension migration guardrails, docs checks, and CI-style safety checks.

Route First

  • Use this sub-skill for make init, backend/frontend test selection, uv sync caveats, formatting, generated component indexes, docs builds, package version updates, migration guards, release checks, and CI helper failures.
  • Route production runbooks, Docker Compose, cloud deployment, storage volumes, runtime secrets, and operational environment variables to ../deployment-and-operations/SKILL.md.
  • Route Python component class design, inputs/outputs, bundles, and component-specific tests to ../component-development/SKILL.md.
  • Route FastAPI route/service/auth/database implementation details to ../backend-runtime/SKILL.md after this sub-skill helps choose checks.
  • Route lfx run, lfx serve, and executor behavior to ../executor-cli/SKILL.md; route SDK client implementation to ../sdk-and-api-clients/SKILL.md.

Maintenance Workflow

  1. Classify the touched area: root package metadata, backend/base, lfx, SDK, frontend, docs, migrations, generated artifacts, or bundle/extension policy.
  2. Read references/maintainer-checks.md for setup, generated-file, migration, versioning, docs, and CI-style guardrails.
  3. Read references/testing-and-formatting.md to choose the smallest safe command set before broad test runs.
  4. Use references/troubleshooting.md when install/imports, optional dependencies, config/schema, CLI/API usage, backend/runtime, or environment boundaries fail.
  5. Run the bundled deprecated-import checker when changing components, bundles, or LangChain import paths:
python scripts/check_deprecated_imports.py
python scripts/check_deprecated_imports.py src/lfx/src/lfx/components src/bundles --format github

Common Command Map

make check_tools
make init
make backend
make frontend
make format_backend
make format_frontend
make format_frontend_check
make unit_tests args="src/backend/tests/unit/services/authorization/test_guards.py -q" async=false
make lfx_test args="tests/unit/cli/test_validate_command.py -q"
make sdk_test args="tests/test_serialization.py -q"
make test_frontend_file src/__tests__/path/to/test.test.tsx
make build_component_index
make alembic-check
make docs_build
make patch v=1.10.2

Prefer focused commands while iterating, then broaden to package-level checks before handoff. Always use uv run or Make targets for Python commands in this workspace.

Safety Rules

  • Do not rename released component classes, mutate persisted flow identifiers, or remove migration-table entries without an explicit compatibility plan.
  • Do not run credentialed provider tests, live external services, GPU/transformer workloads, Docker/cloud deployment, package publishing, or destructive database migration commands unless the user explicitly asks and the environment is safe.
  • Before package-specific tests in src/backend/base, src/lfx, or src/sdk, sync that package's dev dependencies; the root workspace sync may not install all package-local test extras.
  • For generated files, rebuild with the repository command and review the diff; do not hand-edit generated frontend bundles, component indexes, lockfiles, or migration tables unless the generator/guard says so.

Evidence Distilled

This guidance is distilled from Langflow maintainer docs, development setup docs, root and frontend Make targets, Python workspace metadata, package Makefiles, migration and CI helper scripts, backend/LFX/SDK/frontend test layouts, installed package facts, and existing repo-agent instructions. Runtime files here are self-contained; source paths named in prose are evidence names, not required reading for future agents.

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