Generate LLM indicator rules from a piece/section of a Chinese financial document. Reads a section excerpt, asks the LLM to produce rules (indicator, instruction, position, documenttype), validates them with pydantic, and persists to the rules database + this skill's scripts/ dir.
Generate a full end-to-end extraction script for a documenttype, drawing on the LLM and script rules in the rules database. Validates the output with pydantic and saves the runnable script to the skill's scripts/ dir.
Generate LLM indicator rules for a whole PDF by splitting it into chapters via the parsed outline and producing rules per chapter. Validates with pydantic and persists to the rules database + this skill's scripts/ dir.
Read all LLM rules for a documenttype from the rules database and generate a script rule per target indicator via the LLM. Validates with pydantic and persists to the rules database + this skill's scripts/ dir.
Read a single indicator/document's LLM rules from the rules database and generate script rules (extractrule, position, documenttype) for them via the LLM. Validates with pydantic and persists to the rules database + this skill's scripts/ dir.
Chinese financial report MCP server — 31 Shenwan industry AI rule system, outline extraction, AI structured extraction, Elasticsearch store + search, industry rules dashboard. Runs locally from the fd-cn-report Python package.
MCP server "fd-daas-mcp" as configured in FindDataTechnology/fd-daas-mcp. Launched with /Users/chengsishi/code/DAAS/fd-daas-mcp/bin/fd-daas-mcp-server. Needs 1 environment variable to run.
Commercial, login-gated MCP server serving FindData's own crawled data — zero upstream-provider exposure. Runs locally from the fd-find-data-business-mcp Python package.
★not rated 0 29d agoA
tokens not measured
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: