MCP server exposing the Galaxy tool refactoring facade to AI agents.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"galaxy-tool-refactor-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# galaxy-tool-refactor-mcp An **MCP server** that exposes the Galaxy tool refactoring framework to AI coding agents. A tier-4 sibling of `galaxy-tool-refactor-cli`: both wrap the `galaxy-tool-refactor-registry` facade (tier 3.6) for a different audience — the CLI for humans at a terminal, this for agents over the Model Context Protocol. ## Shape The facade is **library-first** (structured args in, structured results out, no disk writes unless asked, introspectable), so the server is a *thin adapter*: - **`service.py`** — protocol-agnostic. Pure functions that take XML as a `str` (plus `rulesets` / `select` / `ignore`) and return JSON-able `dict`s by calling the facade. No `mcp` import; fully unit-tested. - **`server.py`** — the FastMCP binding. A small handler per tool delegates to `service`, and is the error boundary (facade `UnknownRuleset` / `UnknownRuleCode` / `UnknownProfile` and tier-1 `ToolXmlSyntaxError` → a clean MCP tool error). ## Tools | MCP tool | What it does | |---|---| | `format_tool` | Apply a ruleset's fixable rules then format; returns canonical XML + advisory notes. | | `upgrade_tool` | Repair then profile placement then format; minimal-bump by default (`profile=` moves only when strictly needed for validity; `modernize` opts into the behavior-preserving walk, capped at the lower of the behaviour ceiling and the deployment ceiling, the newest profile every major public Galaxy server runs; `allow_behavior_change` lifts the walk's behaviour gate only, `target_profile` caps the walk and may exceed the deployment ceiling); returns upgraded XML, steps applied, the behavior-preserving verdict, `baseline_profile`/`reached_profile`/`stopped_at`/`blocking_codes`/`auto_fixed_codes`, and notes. | | `check_tool` | Report-only detect over the selected rules; returns the findings (each flagged fixable vs advisory). | | `convert_help_tool` | Convert an RST `<help>` body to Markdown when provable (profile ≥ 24.2 + render-equivalence gate); retur…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.