Reorganize task categories from opaque a/b to descriptive names

Replace category_a/category_b directories with algorithm, pipeline,
environment, filesystem, and process. Add separate mode field (solve/convert)
to decouple orchestration from capability grouping. Add per-category
summary and questionnaire breakdowns to both terminal report and HTML export.

tasks/filesystem/file_organizer.toml

@@ -0,0 +1,29 @@
+name = "file_organizer"
+category = "filesystem"
+mode = "solve"
+description = """
+You are given a working directory containing several files with extensions.
+Read a list of extension-to-directory mappings from stdin, one per line, in the format:
+  ext:dirname
+
+For example, "txt:documents" means move all .txt files into a subdirectory called "documents".
+
+Create the target directories if they don't exist, then move matching files into them.
+After processing, print each moved file as "filename -> dirname/filename", sorted alphabetically by filename.
+Files that don't match any mapping should be left in place (don't print anything for them).
+"""
+
+[[test_cases]]
+stdin = """txt:documents
+log:logs"""
+expected_stdout = """app.log -> logs/app.log
+notes.txt -> documents/notes.txt
+readme.txt -> documents/readme.txt"""
+setup_files = { "notes.txt" = "some notes", "readme.txt" = "a readme", "app.log" = "log data", "image.png" = "fake png" }
+expected_files = { "documents/notes.txt" = "some notes", "documents/readme.txt" = "a readme", "logs/app.log" = "log data", "image.png" = "fake png" }
+
+[[test_cases]]
+stdin = "csv:data"
+expected_stdout = """report.csv -> data/report.csv"""
+setup_files = { "report.csv" = "a,b,c", "other.txt" = "hello" }
+expected_files = { "data/report.csv" = "a,b,c", "other.txt" = "hello" }