Add gcache-inspector, binlogsum, redactsql, and innodump to the catalog

Four offline diagnostic/forensic tools by Przemek Malkowski, researched
from their own READMEs. All are single static Go binaries with no
external dependencies; none have a separate project site, so `link`
points straight at the GitHub repo with no redundant `github` field,
matching the convention already used for alisql.md.

- content/resources/software/gcache-inspector.md — offline Galera
  GCache (write-set cache) reader for Percona XtraDB Cluster and
  MariaDB Galera Cluster. Monitoring & Observability + Replication &
  High Availability.
- content/resources/software/binlogsum.md — summarizer for decoded
  MySQL/MariaDB binary logs (per-table writes, GTID ranges,
  writes-over-time histograms). Same two subcategories, for the same
  reason as gcache-inspector — it's a diagnostic tool for replication
  data specifically.
- content/resources/software/redactsql.md — local SQL schema
  anonymizer (MySQL/MariaDB + PostgreSQL, though only the former is a
  tagged compatibility value here). Security + Schema Management &
  Migration.
- content/resources/software/innodump.md — offline InnoDB tablespace
  reader/dumper (.ibd, ibdata1), including deleted-row recovery.
  Backup & Recovery.

No `member`/`last_reviewed`/pricing beyond "Open Source" (all four are
GPL-3.0-or-later per their repos, redactsql is MIT) — same
no-fabricated-content approach as the previous three additions. No
custom logos; auto-generated initials only.

Testing: hugo --minify — clean build, 199 pages, 0 errors. Verified
all four render correctly in both themes and appear in the Software &
Tools listing.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Author: Mazizova

content/resources/software/binlogsum.md

@@ -0,0 +1,19 @@
+---
+title: "binlogsum"
+link: "https://github.com/PrzemekMalkowski/binlogsum"
+description: "A summarizer for decoded MySQL and MariaDB binary logs — turns mysqlbinlog output into per-table write statistics, GTID/transaction ranges, and writes-over-time histograms for diagnosing replication delay and unusual write activity."
+subcategories:
+  - Monitoring & Observability
+  - Replication & High Availability
+compatibility:
+  - MySQL
+  - MariaDB
+deployment:
+  - Self-Hosted
+pricing:
+  - Open Source
+---
+
+binlogsum reads the decoded output of `mysqlbinlog` and reports on what actually happened in a given window of binary log activity: time range covered, server versions and per-server event counts, GTID and transaction ID ranges, a breakdown of query types, and which tables received the most writes and when. It distinguishes tables that were actually updated in a transaction from ones merely referenced, and surfaces the original SQL when `binlog_rows_query_log_events` is enabled — useful for tracing down the cause of a replication lag spike or an unexpected burst of writes.
+
+It supports both ROW and STATEMENT binlog formats and can output a terminal report with Unicode histograms, an interactive web UI with zoomable graphs, or a self-contained HTML snapshot for sharing. Like the author's other tools, it's pure Go with no external dependencies, built as a single static binary.

content/resources/software/gcache-inspector.md

@@ -0,0 +1,20 @@
+---
+title: "gcache-inspector"
+link: "https://github.com/PrzemekMalkowski/gcache-inspector"
+description: "An offline reader for Galera's write-set cache (GCache) files on Percona XtraDB Cluster and MariaDB Galera Cluster, summarizing per-table activity and decoding write-sets into row-level output without a database connection."
+subcategories:
+  - Monitoring & Observability
+  - Replication & High Availability
+compatibility:
+  - MariaDB
+  - Percona Server
+  - Galera
+deployment:
+  - Self-Hosted
+pricing:
+  - Open Source
+---
+
+gcache-inspector parses Galera's GCache write-set cache files directly, without connecting to a running cluster, which makes it useful for diagnosing replication issues on nodes that are stopped, crashed, or otherwise unreachable. It identifies which tables were modified, attributes each write-set to its Galera sequence number, and derives timestamps from the binlog commit data embedded in the cache — producing an activity summary of inserts, updates, deletes, DDL, and bytes per table.
+
+For deeper inspection, it can decode individual write-sets into row-level output comparable to `mysqlbinlog`, with a seqno range selector for targeting specific spans of activity. It auto-detects MySQL v2 vs. MariaDB v1 row event formats rather than assuming a fixed layout, and supports encrypted caches via keyring files or HashiCorp Vault. It's a static Go binary with no external dependencies.

content/resources/software/innodump.md

@@ -0,0 +1,18 @@
+---
+title: "innodump"
+link: "https://github.com/PrzemekMalkowski/innodump"
+description: "An offline reader for InnoDB tablespace files (.ibd, ibdata1) across MySQL 5.6–8.4 and MariaDB — reconstructs CREATE TABLE statements and extracts row data via direct B+tree traversal, without a running database server."
+subcategories:
+  - Backup & Recovery
+compatibility:
+  - MySQL
+  - MariaDB
+deployment:
+  - Self-Hosted
+pricing:
+  - Open Source
+---
+
+innodump reads InnoDB tablespace files — file-per-table `.ibd` files or the shared `ibdata1` — directly off disk, reconstructing `CREATE TABLE` statements from embedded SDI dictionary information or `.frm` files, then walking the B+tree to extract row data. It handles every InnoDB row format (DYNAMIC, COMPACT, REDUNDANT, COMPRESSED) and understands `INSTANT ADD/DROP COLUMN` history, and can recover deleted-but-not-yet-purged rows with a `--deleted-only` flag. Output is either SQL `INSERT` statements or MySQL Shell-compatible TSV, for a single file or recursively across a directory, with a `--skip-corrupted` option to tolerate damaged pages.
+
+It's explicitly read-only and, per the author, "not a backup tool" — it's meant for consistent snapshots (a stopped server, `FLUSH TABLES ... FOR EXPORT`, or an existing backup), offering best-effort data recovery from files that are no longer attached to a live server, not a substitute for one that is. Written in Go with no external dependencies, GPL-3.0-or-later licensed.

content/resources/software/redactsql.md

@@ -0,0 +1,19 @@
+---
+title: "redactsql"
+link: "https://github.com/PrzemekMalkowski/redactsql"
+description: "A local SQL schema anonymizer for MySQL/MariaDB and PostgreSQL — replaces table, column, index, and constraint names with neutral placeholders while preserving structure, so dumps can be shared safely for debugging or bug reports."
+subcategories:
+  - Security
+  - Schema Management & Migration
+compatibility:
+  - MySQL
+  - MariaDB
+deployment:
+  - Self-Hosted
+pricing:
+  - Open Source
+---
+
+redactsql anonymizes a SQL schema — tables, columns, indexes, constraints, sequences, and types — replacing anything that could leak proprietary naming or business context with generic placeholders, while keeping foreign-key relationships and overall schema topology intact. That makes it possible to share a real dump for debugging, documentation, or a public bug report without exposing what a company's schema actually says about its product or data. Well-known system schemas (`public`, `pg_catalog`, `mysql`, `sys`, and similar) are left untouched.
+
+It handles MySQL/MariaDB (backtick quoting, engine-specific syntax) and PostgreSQL (sequences, custom types, dollar-quoted strings) dialects, auto-detecting which one it's looking at or accepting an explicit `-dialect` flag. Everything runs locally — nothing is transmitted or stored externally — either through a local web UI at `127.0.0.1:8585` or a CLI that reads files or pipes from stdin, with options like `-remove-fk` to strip foreign keys and `-keep-id` (on by default) to leave `id` columns untouched.