mcp-scraper 0.56.0 → 0.57.1

package/CHANGELOG.md

@@ -4,6 +4,34 @@ All notable changes to MCP Scraper are documented here. The format is based on [
 
 ## [Unreleased]
 
+## [0.57.1] - 2026-08-16
+
+### Fixed
+
+- Commons image hosting resolves the dedicated public store token and refuses to store anything unless the resolved store is Vercel Blob. `BLOB_READ_WRITE_TOKEN` is empty in production, so the previous release would have written a `file://` path into a public wiki entity instead of failing; hosting now fails loudly with `image_storage_unconfigured`.
+
+### Added
+
+- `scripts/backfill-commons-images.ts` hosts the still-reachable featured images and article media on Commons storage, rewrites the entity, and records a contribution-ledger entry per rewrite. Dry run by default; `--apply` to write. Run against production: 15 featured images and 20 media sets moved onto Commons hosting; 26 sources were already dead 404s and one SVG is rejected by policy.
+
+## [0.57.0] - 2026-08-16
+
+### Fixed
+
+- Transparent Commons now hosts its own images. Entities stored whatever third-party URL an author supplied and the public wiki hotlinked it, so 26 of 41 live pages had a broken featured image once the source site rotated its uploads. `commons_submit_entity` downloads external images, validates them, stores them permanently, and rewrites the entity to the hosted copy, so a foreign image URL can no longer be persisted.
+- Commons search no longer matches a term inside an unrelated word or ranks purely by recency. Searching `CIA` returned eight physical-therapy pages — matching inside `fascia`, `myofascial`, and `special` — while missing the article with CIA in its title. Terms now match on word boundaries and results are scored title > tags/keywords > description > governed claims > body before the recency tiebreak.
+- See Also concepts resolve to pages whose slug was shortened by hand, not only to pages whose slug matches their title.
+
+### Added
+
+- `commons_host_image` hosts an image from a public https URL or base64 bytes and returns the permanent Commons URL. JPEG, PNG, GIF, and WebP up to 10 MB; SVG is rejected because the wiki serves images to the public web. Identical bytes are stored once and reused. URL fetches enforce https, reject private, loopback, link-local, and carrier-grade NAT addresses, cap redirects, and verify the real bytes rather than the declared content type.
+
+## [0.56.1] - 2026-08-16
+
+### Fixed
+
+- See Also concepts now link to pages written after the mention. A related link stored with `needsLink: true` is resolved against published entities by normalized title on every read, so `commons_get_entity`, `commons_search_entities`, and the public `/wiki/` projection return a real slug instead of an unresolved concept. 37 dangling mentions across 16 concepts linked immediately, and `commons_list_needs_links` now reports only concepts that genuinely have no page.
+
 ## [0.56.0] - 2026-08-16
 
 ### Fixed

package/README.md

@@ -90,7 +90,7 @@ Build the branded one-click bundle:
 npm run build:mcpb
 ```
 
-The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.56.0`, SHA-256 `72928532db3280859d5d51283b1828acb66911cfb7cbd179ffc7925fd33d0d73`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
+The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.57.1`, SHA-256 `ea84195321a47e7550348be6f61ed3e6a293a679c20a6cd39e77cd37870bd636`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
 
 The MCPB install exposes every tool — web-intelligence plus all `browser_*` tools — through the one `mcp-scraper` server.
 
@@ -248,7 +248,7 @@ The `mcp-scraper` server (and the MCPB bundle, which runs it) exposes both secti
 
 All MCP tools expose output schemas and return `structuredContent` with the IDs, URLs, CSV paths, transcripts, browser session handles, replay paths, artifacts, recipe fields, or blueprint fields needed by the next step. Browser Agent tools keep a JSON text block for older clients, but structured data is the primary contract. All tools carry MCP annotations; file-writing tools such as replay downloads and annotations state their filesystem side effects.
 
-The canonical tool inventory is generated at `docs/mcp-tool-manifest.generated.json`. The unified server exposes 250 tools: 149 scraper, browser, workflow, billing, and connected-service tools plus 101 durable-memory tools. The scraper-side inventory includes governed Local Sourcebook tools, Transparent Commons entity, linkset, and publication tools, the searchable News Publisher and Editorial Reading Room template workflow, and 24 X-Ray tools for pixel administration, reporting, attribution journeys, campaign links, exports, and scheduled reports. Successful evidence-compiled Local Sourcebook revisions publish automatically to their canonical `localsourcebook.com` category profile and review URLs; administrator controls handle exceptional rejection or unpublishing. Release verification compares the exact local and hosted tool-name sets, not only the count.
+The canonical tool inventory is generated at `docs/mcp-tool-manifest.generated.json`. The unified server exposes 251 tools: 150 scraper, browser, workflow, billing, and connected-service tools plus 101 durable-memory tools. The scraper-side inventory includes governed Local Sourcebook tools, Transparent Commons entity, linkset, and publication tools, the searchable News Publisher and Editorial Reading Room template workflow, and 24 X-Ray tools for pixel administration, reporting, attribution journeys, campaign links, exports, and scheduled reports. Successful evidence-compiled Local Sourcebook revisions publish automatically to their canonical `localsourcebook.com` category profile and review URLs; administrator controls handle exceptional rejection or unpublishing. Release verification compares the exact local and hosted tool-name sets, not only the count.
 
 For contract parity, stdio and MCPB memory calls invoke the matching public tool on the hosted MCP Scraper `/mcp` endpoint. The hosted aggregate runtime owns MCP Scraper-specific billing, scheduling, credential, and in-process cutover policy; its internal `/memory/mcp-call` bridge is a fallback to the standalone memory service, not the public stdio execution path. Direct `mcp-memory` OAuth and stdio clients continue to use `memory.mcpscraper.dev` and must be verified as a separate dependent release surface.