diff --git a/src/projspec/proj/__init__.py b/src/projspec/proj/__init__.py
index 80dbc40..b52d2da 100644
--- a/src/projspec/proj/__init__.py
+++ b/src/projspec/proj/__init__.py
@@ -3,6 +3,7 @@
 from projspec.proj.base import ParseFailed, Project, ProjectSpec, ProjectExtra
 
 from projspec.proj.ai import AIEnabled
+from projspec.proj.backstage import BackstageCatalog
 from projspec.proj.briefcase import Briefcase
 from projspec.proj.conda_package import CondaRecipe, RattlerRecipe
 from projspec.proj.conda_project import CondaProject
@@ -29,6 +30,7 @@
     "Project",
     "ProjectSpec",
     "AIEnabled",
+    "BackstageCatalog",
     "Briefcase",
     "Cited",
     "Zenodo",
diff --git a/src/projspec/proj/backstage.py b/src/projspec/proj/backstage.py
index 55b7d33..d451195 100644
--- a/src/projspec/proj/backstage.py
+++ b/src/projspec/proj/backstage.py
@@ -1,13 +1,98 @@
+import os
+
+import yaml
+
 from projspec.proj import ProjectSpec
+from projspec.proj.base import ParseFailed
+from projspec.utils import AttrDict
 
 
 class BackstageCatalog(ProjectSpec):
+    """A Backstage software catalog descriptor file (catalog-info.yaml).
+
+    Backstage catalogs describe software components, APIs, resources, systems,
+    domains and other entities using a Kubernetes-style envelope with ``apiVersion``,
+    ``kind``, ``metadata``, and ``spec`` sections.  A single file may contain multiple
+    ``---``-separated documents.
+    """
+
     spec_doc = "https://backstage.io/docs/features/software-catalog/descriptor-format/"
 
     def match(self) -> bool:
         return "catalog-info.yaml" in self.proj.basenames
 
-    # spec is "---"-separated list of entries, each of which has a "kind", and a "spec"
-    # as well as general metadata. The spec normally has a "type" to tell you what we're
-    # really talking about. Each entry should have "apiVersion: backstage.io/*"
-    # where v1alpha1 is the only known version when this was written.
+    def parse(self) -> None:
+        from projspec.content.metadata import DescriptiveMetadata
+
+        try:
+            with self.proj.get_file("catalog-info.yaml") as f:
+                raw = f.read()
+        except OSError as exc:
+            raise ParseFailed(f"Could not read catalog-info.yaml: {exc}") from exc
+
+        # The file may contain multiple YAML documents separated by "---"
+        try:
+            docs = list(yaml.safe_load_all(raw))
+        except yaml.YAMLError as exc:
+            raise ParseFailed(f"Invalid YAML in catalog-info.yaml: {exc}") from exc
+
+        docs = [d for d in docs if isinstance(d, dict)]
+        if not docs:
+            raise ParseFailed("catalog-info.yaml contains no valid entity documents")
+
+        # Only accept documents that declare a backstage apiVersion
+        backstage_docs = [
+            d for d in docs if str(d.get("apiVersion", "")).startswith("backstage.io/")
+        ]
+        if not backstage_docs:
+            raise ParseFailed(
+                "catalog-info.yaml contains no backstage.io entity documents"
+            )
+
+        # Build one DescriptiveMetadata per entity, keyed by "<kind>.<name>".
+        meta_entries = AttrDict()
+        for doc in backstage_docs:
+            kind = doc.get("kind", "unknown")
+            metadata = (
+                doc.get("metadata", {}) if isinstance(doc.get("metadata"), dict) else {}
+            )
+            spec = doc.get("spec", {}) if isinstance(doc.get("spec"), dict) else {}
+
+            name = metadata.get("name", "unnamed")
+            key = f"{kind.lower()}.{name}"
+
+            entry: dict[str, str] = {"kind": kind}
+            for field in ("name", "title", "description", "namespace"):
+                if val := metadata.get(field):
+                    entry[field] = str(val)
+            for field in ("type", "lifecycle", "owner"):
+                if val := spec.get(field):
+                    entry[field] = str(val)
+            tags = metadata.get("tags", [])
+            if tags:
+                entry["tags"] = ", ".join(str(t) for t in tags)
+
+            meta_entries[key] = DescriptiveMetadata(proj=self.proj, meta=entry)
+
+        self._contents = AttrDict(descriptive_metadata=meta_entries)
+        self._artifacts = AttrDict()
+
+    @staticmethod
+    def _create(path: str) -> None:
+        """Scaffold a minimal Backstage catalog-info.yaml."""
+        name = os.path.basename(path)
+        doc = {
+            "apiVersion": "backstage.io/v1alpha1",
+            "kind": "Component",
+            "metadata": {
+                "name": name,
+                "description": f"A {name} component",
+            },
+            "spec": {
+                "type": "service",
+                "lifecycle": "experimental",
+                "owner": "team-default",
+            },
+        }
+        with open(f"{path}/catalog-info.yaml", "wt") as f:
+            yaml.dump(doc, f, default_flow_style=False, sort_keys=False)
diff --git a/src/projspec/proj/documentation.py b/src/projspec/proj/documentation.py
index 9273309..e6fdeb1 100644
--- a/src/projspec/proj/documentation.py
+++ b/src/projspec/proj/documentation.py
@@ -1,6 +1,11 @@
+import os
 import re
 
+import toml
+
 from projspec.proj import ProjectSpec
+from projspec.proj.base import ParseFailed
+from projspec.utils import AttrDict, PickleableTomlDecoder
 
 
 class MDBook(ProjectSpec):
@@ -18,6 +23,83 @@ class MDBook(ProjectSpec):
     def match(self) -> bool:
         return "book.toml" in self.proj.basenames
 
+    def parse(self) -> None:
+        from projspec.artifact.base import FileArtifact
+        from projspec.artifact.process import Server
+        from projspec.content.metadata import DescriptiveMetadata
+
+        try:
+            with self.proj.get_file("book.toml", text=False) as f:
+                cfg = toml.loads(f.read().decode(), decoder=PickleableTomlDecoder())
+        except (OSError, toml.TomlDecodeError) as exc:
+            raise ParseFailed(f"Could not read book.toml: {exc}") from exc
+
+        book = cfg.get("book", {})
+
+        # ------------------------------------------------------------------ #
+        # Contents — descriptive metadata from the [book] table
+        # ------------------------------------------------------------------ #
+        meta: dict[str, str] = {}
+        for key in ("title", "description", "language"):
+            if val := book.get(key):
+                meta[key] = str(val)
+        authors = book.get("authors", [])
+        if authors:
+            meta["authors"] = ", ".join(authors)
+
+        self._contents = AttrDict(
+            descriptive_metadata=DescriptiveMetadata(proj=self.proj, meta=meta)
+        )
+
+        # ------------------------------------------------------------------ #
+        # Artifacts
+        # ------------------------------------------------------------------ #
+        # build-dir defaults to "book/" relative to the book root
+        build_dir = cfg.get("build", {}).get("build-dir", "book")
+        if not build_dir.startswith("/"):
+            build_dir = f"{self.proj.url}/{build_dir}"
+
+        arts = AttrDict()
+        # mdbook build → produces static HTML in build-dir
+        arts["book"] = FileArtifact(
+            proj=self.proj,
+            cmd=["mdbook", "build"],
+            fn=f"{build_dir}/index.html",
+        )
+        # mdbook serve → live-reloading local server
+        arts["server"] = Server(
+            proj=self.proj,
+            cmd=["mdbook", "serve"],
+        )
+
+        self._artifacts = arts
+
+    @staticmethod
+    def _create(path: str) -> None:
+        """Scaffold a minimal but valid mdBook project."""
+        name = os.path.basename(path)
+
+        # book.toml — required configuration file
+        with open(f"{path}/book.toml", "wt") as f:
+            f.write(
+                f"[book]\n"
+                f'title = "{name}"\n'
+                f"authors = []\n"
+                f'description = ""\n'
+                f"\n"
+                f"[build]\n"
+                f'build-dir = "book"\n'
+            )
+
+        # src/SUMMARY.md — required entry point
+        os.makedirs(f"{path}/src", exist_ok=True)
+        with open(f"{path}/src/SUMMARY.md", "wt") as f:
+            f.write(f"# Summary\n\n- [Introduction](./introduction.md)\n")
+
+        # src/introduction.md — first chapter
+        with open(f"{path}/src/introduction.md", "wt") as f:
+            f.write(f"# Introduction\n\nWelcome to {name}.\n")
+
 
 class RTD(ProjectSpec):
     """Documentation to be processes by ReadTheDocs
@@ -34,12 +116,146 @@ def match(self) -> bool:
         return any(re.match("[.]?readthedocs.y[a]?ml", _) for _ in self.proj.basenames)
 
     def parse(self) -> None:
-        # supports mkdocs and sphinx builders
-        # build env usually in `python.install[*].requirements`, which can
-        # point to a requirements.txt or environment.yaml for conda env.
-
-        # Artifact of HTML output. Classically with `make html` (in docs/ unless otherwise
-        # specified), and the output goes into docs/build/html unless the config.py file
-        # says different. RTD actually does
-        # > python -m sphinx -T -W --keep-going -b html -d _build/doctrees -D language=en . $READTHEDOCS_OUTPUT/html
-        pass
+        import yaml
+
+        from projspec.artifact.base import FileArtifact
+        from projspec.content.environment import Environment, Precision, Stack
+
+        # Locate and read the config file
+        cfg_name = next(
+            _ for _ in self.proj.basenames if re.match("[.]?readthedocs.y[a]?ml", _)
+        )
+        try:
+            with self.proj.get_file(cfg_name) as f:
+                cfg = yaml.safe_load(f)
+        except (OSError, yaml.YAMLError) as exc:
+            raise ParseFailed(f"Could not read {cfg_name}: {exc}") from exc
+
+        if not isinstance(cfg, dict):
+            raise ParseFailed(f"{cfg_name} did not parse to a mapping")
+
+        conts = AttrDict()
+        arts = AttrDict()
+
+        # ------------------------------------------------------------------ #
+        # Environment — conda env file or pip requirements files
+        # ------------------------------------------------------------------ #
+        conda_env_path = cfg.get("conda", {}).get("environment")
+        if conda_env_path:
+            try:
+                with self.proj.fs.open(f"{self.proj.url}/{conda_env_path}", "rt") as f:
+                    env_data = yaml.safe_load(f)
+                conts["environment"] = AttrDict(
+                    default=Environment(
+                        proj=self.proj,
+                        stack=Stack.CONDA,
+                        precision=Precision.SPEC,
+                        packages=env_data.get("dependencies", []),
+                        channels=env_data.get("channels", []),
+                    )
+                )
+            except (OSError, yaml.YAMLError):
+                pass
+        else:
+            # Collect requirements files listed under python.install[*].requirements
+            req_packages: list[str] = []
+            for install in cfg.get("python", {}).get("install", []):
+                req_path = install.get("requirements")
+                if not req_path:
+                    continue
+                try:
+                    with self.proj.get_file(req_path, text=False) as f:
+                        lines = [
+                            ln.strip()
+                            for ln in f.read().decode().splitlines()
+                            if ln.strip() and not ln.strip().startswith("#")
+                        ]
+                    req_packages.extend(lines)
+                except OSError:
+                    pass
+            if req_packages:
+                # Add python version constraint if declared
+                py_ver = cfg.get("build", {}).get("tools", {}).get("python")
+                if py_ver:
+                    req_packages.append(f"python =={py_ver}.*")
+                precision = (
+                    Precision.LOCK
+                    if all(
+                        "==" in p for p in req_packages if not p.startswith("python")
+                    )
+                    else Precision.SPEC
+                )
+                conts["environment"] = AttrDict(
+                    default=Environment(
+                        proj=self.proj,
+                        stack=Stack.PIP,
+                        precision=precision,
+                        packages=req_packages,
+                    )
+                )
+
+        # ------------------------------------------------------------------ #
+        # Artifacts — sphinx or mkdocs build process
+        # ------------------------------------------------------------------ #
+        if "sphinx" in cfg:
+            conf_py = cfg["sphinx"].get("configuration", "docs/conf.py")
+            docs_dir = conf_py.rsplit("/", 1)[0] if "/" in conf_py else "."
+            arts["docs"] = FileArtifact(
+                proj=self.proj,
+                cmd=[
+                    "sphinx-build",
+                    "-b",
+                    "html",
+                    docs_dir,
+                    f"{docs_dir}/_build/html",
+                ],
+                fn=f"{self.proj.url}/{docs_dir}/_build/html/index.html",
+            )
+        elif "mkdocs" in cfg:
+            arts["docs"] = FileArtifact(
+                proj=self.proj,
+                cmd=["mkdocs", "build"],
+                fn=f"{self.proj.url}/site/index.html",
+            )
+
+        self._contents = conts
+        self._artifacts = arts
+
+    @staticmethod
+    def _create(path: str) -> None:
+        """Scaffold a minimal RTD project using Sphinx."""
+        # .readthedocs.yaml — RTD configuration
+        with open(f"{path}/.readthedocs.yaml", "wt") as f:
+            f.write(
+                "version: 2\n"
+                "\n"
+                "build:\n"
+                "  os: ubuntu-24.04\n"
+                "  tools:\n"
+                '    python: "3.12"\n'
+                "\n"
+                "sphinx:\n"
+                "  configuration: docs/conf.py\n"
+                "\n"
+                "python:\n"
+                "  install:\n"
+                "    - requirements: docs/requirements.txt\n"
+            )
+
+        # docs/conf.py — minimal Sphinx configuration
+        os.makedirs(f"{path}/docs", exist_ok=True)
+        name = os.path.basename(path)
+        with open(f"{path}/docs/conf.py", "wt") as f:
+            f.write(
+                f'project = "{name}"\n'
+                f"extensions = []\n"
+                f'html_theme = "alabaster"\n'
+            )
+
+        # docs/index.rst — root document
+        with open(f"{path}/docs/index.rst", "wt") as f:
+            f.write(f"{name}\n{'=' * len(name)}\n\n.. toctree::\n   :maxdepth: 2\n")
+
+        # docs/requirements.txt — build dependencies
+        with open(f"{path}/docs/requirements.txt", "wt") as f:
+            f.write("sphinx\n")
diff --git a/src/projspec/proj/golang.py b/src/projspec/proj/golang.py
index 31bcaf5..a65340c 100644
--- a/src/projspec/proj/golang.py
+++ b/src/projspec/proj/golang.py
@@ -1,8 +1,12 @@
+import re
+
 from projspec.proj import ProjectSpec
+from projspec.proj.base import ParseFailed
+from projspec.utils import AttrDict
 
 
 class Golang(ProjectSpec):
-    """A Go language module project"""
+    """A Go module project, identified by the presence of a go.mod file."""
 
     spec_doc = "https://go.dev/doc/modules/gomod-ref"
 
@@ -10,9 +14,87 @@ def match(self) -> bool:
         return "go.mod" in self.proj.basenames
 
     def parse(self) -> None:
-        # TODO: probably `go run` or `go build` artifacts
-        # dependencies via the require keyword
-        pass
+        from projspec.artifact.base import FileArtifact
+        from projspec.artifact.process import Process
+        from projspec.content.metadata import DescriptiveMetadata
+
+        try:
+            with self.proj.fs.open(self.proj.basenames["go.mod"], "rb") as f:
+                text = f.read().decode()
+        except OSError as exc:
+            raise ParseFailed(f"Could not read go.mod: {exc}") from exc
+
+        # ------------------------------------------------------------------ #
+        # Parse go.mod directives with simple regex — avoids a Go dependency
+        # ------------------------------------------------------------------ #
+        # module path: "module example.com/mymodule"
+        module_match = re.search(r"^module\s+(\S+)", text, re.MULTILINE)
+        module_path = module_match.group(1) if module_match else ""
+
+        # minimum Go version: "go 1.21"
+        go_ver_match = re.search(r"^go\s+(\S+)", text, re.MULTILINE)
+        go_version = go_ver_match.group(1) if go_ver_match else ""
+
+        # require directives — both single-line and block forms
+        #   require example.com/foo v1.2.3
+        #   require ( example.com/foo v1.2.3 // indirect )
+        requires: list[str] = re.findall(
+            r"^\s+?(\S+)\s+(\S+?)(?:\s*//.*)?$",
+            # extract the body of require blocks
+            "\n".join(
+                "\n".join(block.splitlines())
+                for block in re.findall(
+                    r"^require\s*\(([^)]*)\)", text, re.MULTILINE | re.DOTALL
+                )
+            ),
+            re.MULTILINE,
+        )
+        # Also catch single-line: "require example.com/foo v1.2.3"
+        single_requires: list[tuple[str, str]] = re.findall(
+            r"^require\s+(\S+)\s+(\S+)", text, re.MULTILINE
+        )
+        all_deps = [f"{mod} {ver}" for mod, ver in (requires + single_requires)]
+
+        # ------------------------------------------------------------------ #
+        # Contents
+        # ------------------------------------------------------------------ #
+        meta: dict[str, str] = {}
+        if module_path:
+            meta["module"] = module_path
+        if go_version:
+            meta["go"] = go_version
+
+        self._contents = AttrDict(
+            descriptive_metadata=DescriptiveMetadata(proj=self.proj, meta=meta)
+        )
+
+        # ------------------------------------------------------------------ #
+        # Artifacts
+        # ------------------------------------------------------------------ #
+        arts = AttrDict()
+
+        # go build ./... — compiles all packages; output depends on module layout
+        arts["build"] = Process(
+            proj=self.proj,
+            cmd=["go", "build", "./..."],
+        )
+
+        # go test ./... — runs all tests
+        arts["test"] = Process(
+            proj=self.proj,
+            cmd=["go", "test", "./..."],
+        )
+
+        # If there is a cmd/ subdirectory the convention is one binary per sub-package.
+        # We model the whole tree as a single FileArtifact since we don't walk the tree.
+        if self.proj.fs.isdir(f"{self.proj.url}/cmd"):
+            arts["binary"] = FileArtifact(
+                proj=self.proj,
+                cmd=["go", "build", "-o", "bin/", "./cmd/..."],
+                fn=f"{self.proj.url}/bin/*",
+            )
+
+        self._artifacts = arts
 
     @staticmethod
     def _create(path: str) -> None:
diff --git a/tests/test_roundtrips.py b/tests/test_roundtrips.py
index 9e0e287..48d4985 100644
--- a/tests/test_roundtrips.py
+++ b/tests/test_roundtrips.py
@@ -33,6 +33,9 @@
         "briefcase",
         "conda_project",
         "helm_chart",
+        "MDBook",
+        "RTD",
+        "BackstageCatalog",
     ],
 )
 def test_compliant(tmpdir, cls_name):