v2.0.0 WordNet Synonyms, MCP Registry Publishing & Breaking Rename

[2.0.0] - 2026-08-02

Added

• New word_synonym read-only tool backed by the biz-dfch-asdste100nlp
  library's Nlp class: looks up WordNet synsets for a word and cross-
  references their lemma names against the ASD-STE100 Issue 9 vocabulary,
  returning approved/rejected entries that are synonyms of the input word.
  Registered in tools/words/word_synonym.py, following the same
  structure as word_find/word_fuzzy; the shared Nlp instance is
  created in server.py's lifespan (_nlp/_get_nlp()), wrapping the
  same Vocab instance used by the other vocabulary tools.
• Runtime dependency on biz-dfch-asdste100nlp>=0.1.0 in pyproject.toml.
• tests/tools/words/test_word_synonym.py — unit tests for the
  Nlp.synonym lookup and the word_synonym tool function.
• MCP Registry publishing support via new publish-to-mcp-registry GitHub Actions job:
  • Automatically publishes server metadata to the official MCP Registry
  • Uses GitHub OIDC authentication (zero secrets, consistent with PyPI Trusted Publishing)
  • Auto-patches server.json version fields from git tag at publish time
  • Runs after PyPI publication succeeds in the CI workflow
• server.json at repository root defining MCP Registry metadata:
  • Server identity: io.github.dfch/biz-dfch-asdste100mcp (GitHub OIDC namespace)
  • PyPI package configuration with biz-dfch-asdste100mcp identifier
  • Documentation of all 8 ASDSTE100_MCP_* environment variables with format hints
  • Conforms to MCP Registry schema v2025-12-11
• PyPI ownership verification marker in README.md long-description for registry discovery

Changed

• Updated release workflow documentation in README.md to include MCP Registry publishing step
• Added "MCP Registry Publishing" section in README.md with setup and verification instructions
• Breaking: Upgraded MCP dependency to version 2.0.0, requiring MCP protocol 2.0 or newer
• Breaking: Renamed the word_similar vocabulary tool to word_fuzzy for clarity
  (tools/words/word_similar.py → tools/words/word_fuzzy.py). Any caller invoking the
  tool by its previous name must switch to word_fuzzy.
• Breaking: Renamed the console script entry point from ste100-mcp to
  asdste100-mcp ([project.scripts] in pyproject.toml), and the MCP
  server/Typer app name to match. Update any launch command, OpenCode
  mcp config, or shell alias that invokes ste100-mcp to use
  asdste100-mcp instead.
• Breaking: Renamed all STE100_MCP_* environment variables to
  ASDSTE100_MCP_* (STE100_MCP_TRANSPORT, STE100_MCP_HOST,
  STE100_MCP_PORT, STE100_MCP_FILES, STE100_MCP_USE_STE100,
  STE100_MCP_USE_STE100_TECHNICAL_WORDS, STE100_MCP_RULES_FILES,
  STE100_MCP_USE_STE100_RULES), matching the asdste100-mcp entry point
  name (settings.py's env_prefix, cli.py's envvar= bindings, and
  server.json). Update any .env file, shell environment, or deployment
  config that sets these variables under their previous names.

Full Changelog: v1.0.0...v2.0.0