provider-connectors
AgentScope provider, embedding, formatter, and TTS configuration workflows.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- provider-connectors
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- AgentScope provider, embedding, formatter, and TTS configuration workflows.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/embedding-overview.md1.8 KB
- 参考文档references/model-overview.md2.7 KB
- 参考文档references/troubleshooting.md2.5 KB
- 参考文档references/tts-overview.md1.4 KB
- 脚本scripts/provider_matrix.py3.5 KB
03 SKILL.md 正文
provider-connectors
Use this sub-skill when the task is about provider classes, credentials, provider extras, model names, embeddings, formatters, or TTS.
Read first
references/model-overview.mdfor chat model families and credential shapes.references/embedding-overview.mdfor embedding model families and dimension notes.references/tts-overview.mdfor TTS constructors and streaming defaults.references/troubleshooting.mdfor missing-extra, credential, and provider-specific failures.scripts/provider_matrix.pyfor a safe availability/import check.
Typical triggers
- Choose the right extra for OpenAI, Anthropic, DashScope, Gemini, Ollama, Moonshot, DeepSeek, or XAI.
- Compare chat, embedding, formatter, or TTS constructor defaults.
- Fix a provider import or credential problem.
- Check which providers are available in the current environment.
What belongs here
agentscope.credentialagentscope.modelagentscope.embeddingagentscope.formatteragentscope.tts- provider-specific configuration, env vars, and model names
What does not belong here
- Agent/tool/permission basics →
agent-core - RAG and memory workflows →
rag-memory - Service deployment and API bootstrap →
service-platform - Workspace or sandbox backend setup →
workspace-sandboxes
Use pattern
- Identify the provider family first.
- Read the matching overview reference for the constructor and default values.
- Check whether the provider needs a dedicated extra or only the base package.
- Use
scripts/provider_matrix.pywhen you need a safe current-environment status check. - Escalate to
rag-memoryonly if the issue is really an embedding dimension or retrieval workflow.
Shared diagnostics
- Run
../../scripts/check_env.pyif the entire package looks stale. - Read
references/troubleshooting.mdbefore changing code if the problem is a missing extra, credential class, or model name.
Cross-links
- If the task needs the agent/tool layer, switch to
agent-core. - If the task needs RAG or memory, switch to
rag-memory. - If the task needs service bootstrapping or deployment, switch to
service-platform.
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).