AGENT SKILL

mcp-tools

Operate DeepKE's local MCP service wrapper for NER, RE, AE, and EE extraction tools without relying on source checkout documentation.

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

01 前言 Front Matter

name
mcp-tools
license
MIT
metadata
{ "disco-role": "operating" }
description
Operate DeepKE's local MCP service wrapper for NER, RE, AE, and EE extraction tools without relying on source checkout documentation.
disable-model-invocation
true

02 资源文件 5

03 SKILL.md 正文

DeepKE MCP Tools

Use this sub-skill when the task is to expose DeepKE extraction through an MCP-compatible local server, configure a Cursor/Cline-like MCP client entry, understand the deepke_ner, deepke_re, deepke_ae, or deepke_ee tool arguments, or diagnose environment problems around DEEPKE_PATH, CONDA_PY, CONDA_EE_PY, API-client variables, and mcp.server.fastmcp imports.

Do not use this sub-skill as proof that an extraction model is available. The MCP wrapper only shells out to local DeepKE example predictors; trained checkpoints, compatible example configs, and a trusted local checkout are still required for real predictions.

Quick route

  1. For deployment and client/server wiring, read references/workflows.md.
  2. For exact tool signatures, task-mode limits, return text, and side effects, read references/tool-api-reference.md.
  3. For failures around missing variables, mcp versions, absent checkpoints, source config mutation, client API errors, or path handling, read references/troubleshooting.md.
  4. Before launching a server, run scripts/check_mcp_env.py to verify imports and environment-variable presence without starting the service or calling models.
  5. When only the event-extraction TSV/raw input conversion is needed, run scripts/convert_text_to_tsv.py; it is a standalone adaptation of the pure conversion helper and does not import DeepKE.

Operating guardrails

  • Treat the local MCP server as a wrapper around mutable local DeepKE example directories, not as a stateless API.
  • The original server edits prediction/training YAML files and event-extraction data files before shelling out. Only run it when the downstream task explicitly asks to operate a local DeepKE checkout and the user accepts those mutation risks.
  • Keep all path and credential values in environment variables or client settings; do not write private paths, API keys, or local environment names into generated artifacts.
  • Prefer the bundled diagnostic and conversion scripts for safe checks. They do not launch the MCP server, invoke LLM APIs, run DeepKE predictors, or mutate DeepKE configs.

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