diff --git a/Cargo.lock b/Cargo.lock
index efa04004..b4736427 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -4670,6 +4670,7 @@ dependencies = [
 name = "rss"
 version = "0.1.0"
 dependencies = [
+ "chrono",
  "domain",
  "rss 2.0.13",
 ]
diff --git a/crates/adapters/rss/Cargo.toml b/crates/adapters/rss/Cargo.toml
index bde6130f..2de1ac45 100644
--- a/crates/adapters/rss/Cargo.toml
+++ b/crates/adapters/rss/Cargo.toml
@@ -4,5 +4,8 @@ version = "0.1.0"
 edition = "2024"
 
 [dependencies]
-rss-feed = { package = "rss", version = "2" }
+rss-feed = { package = "rss", version = "2", features = ["atom"] }
 domain = { workspace = true }
+
+[dev-dependencies]
+chrono = { workspace = true }
diff --git a/crates/adapters/rss/src/lib.rs b/crates/adapters/rss/src/lib.rs
index 180ba269..d11d5ff8 100644
--- a/crates/adapters/rss/src/lib.rs
+++ b/crates/adapters/rss/src/lib.rs
@@ -1,19 +1,30 @@
 use domain::models::DiaryEntry;
-use domain::ports::RssFeedRenderer;
+use domain::ports::{FeedChannel, RssFeedRenderer};
+use rss_feed::extension::atom::{self, AtomExtensionBuilder, Link};
 use rss_feed::{ChannelBuilder, GuidBuilder, ItemBuilder};
 
 pub struct RssAdapter {
-    feed_link: String,
+    base_url: String,
 }
 
 impl RssAdapter {
-    pub fn new(feed_link: String) -> Self {
-        Self { feed_link }
+    pub fn new(base_url: String) -> Self {
+        Self {
+            base_url: base_url.trim_end_matches('/').to_string(),
+        }
+    }
+
+    fn absolute(&self, path: &str) -> String {
+        format!("{}{}", self.base_url, path)
     }
 }
 
 impl RssFeedRenderer for RssAdapter {
-    fn render_feed(&self, entries: &[DiaryEntry], title: &str) -> Result<String, String> {
+    fn render_feed(
+        &self,
+        entries: &[DiaryEntry],
+        channel: &FeedChannel<'_>,
+    ) -> Result<String, String> {
         let items = entries
             .iter()
             .map(|e| {
@@ -37,12 +48,14 @@ impl RssFeedRenderer for RssAdapter {
                     .and_utc()
                     .format("%a, %d %b %Y %H:%M:%S +0000")
                     .to_string();
+                let link = self.absolute(&format!("/movies/{}", e.movie().id().value()));
                 let guid = GuidBuilder::default()
                     .value(e.review().id().value().to_string())
                     .permalink(false)
                     .build();
                 ItemBuilder::default()
                     .title(Some(item_title))
+                    .link(Some(link))
                     .description(Some(description))
                     .pub_date(Some(pub_date))
                     .guid(Some(guid))
@@ -50,10 +63,24 @@ impl RssFeedRenderer for RssAdapter {
             })
             .collect::<Vec<_>>();
 
+        let self_link = Link {
+            href: self.absolute(channel.self_path),
+            rel: "self".to_string(),
+            mime_type: Some("application/rss+xml".to_string()),
+            ..Default::default()
+        };
+
         let channel = ChannelBuilder::default()
-            .title(title.to_string())
-            .link(self.feed_link.clone())
-            .description(title.to_string())
+            .title(channel.title.to_string())
+            .link(self.absolute(channel.link_path))
+            .description(channel.title.to_string())
+            .language(Some("en".to_string()))
+            .namespace(("atom".to_string(), atom::NAMESPACE.to_string()))
+            .atom_ext(Some(
+                AtomExtensionBuilder::default()
+                    .links(vec![self_link])
+                    .build(),
+            ))
             .items(items)
             .build();
 
diff --git a/crates/adapters/rss/src/tests/lib.rs b/crates/adapters/rss/src/tests/lib.rs
index fd2c9cbb..0fe98bf4 100644
--- a/crates/adapters/rss/src/tests/lib.rs
+++ b/crates/adapters/rss/src/tests/lib.rs
@@ -1,15 +1,79 @@
 use super::*;
+use domain::models::{Movie, Review};
+use domain::value_objects::{Comment, MovieTitle, Rating, ReleaseYear, UserId};
+
+fn channel<'a>(title: &'a str) -> FeedChannel<'a> {
+    FeedChannel {
+        title,
+        link_path: "/users/abc",
+        self_path: "/users/abc/feed.rss",
+    }
+}
+
+fn entry(spoilers: bool) -> DiaryEntry {
+    let movie = Movie::new(
+        None,
+        MovieTitle::new("Blade Runner".into()).unwrap(),
+        ReleaseYear::new(1982).unwrap(),
+        None,
+        None,
+    );
+    let review = Review::new(
+        movie.id().clone(),
+        UserId::generate(),
+        Rating::new(4).unwrap(),
+        Some(Comment::new("replicants".into()).unwrap()),
+        chrono::Utc::now().naive_utc(),
+        None,
+        spoilers,
+    )
+    .unwrap();
+    DiaryEntry::new(movie, review)
+}
 
 #[test]
 fn render_feed_uses_provided_title() {
     let adapter = RssAdapter::new("http://example.com".into());
-    let xml = adapter.render_feed(&[], "Custom Title").unwrap();
+    let xml = adapter.render_feed(&[], &channel("Custom Title")).unwrap();
     assert!(xml.contains("<title>Custom Title</title>"));
 }
 
 #[test]
 fn render_feed_empty_entries_produces_valid_xml() {
     let adapter = RssAdapter::new("http://example.com".into());
-    let xml = adapter.render_feed(&[], "My Feed").unwrap();
+    let xml = adapter.render_feed(&[], &channel("My Feed")).unwrap();
     assert!(xml.starts_with("<?xml") || xml.starts_with("<rss"));
 }
+
+#[test]
+fn channel_links_are_absolute_with_atom_self() {
+    let adapter = RssAdapter::new("http://example.com/".into());
+    let xml = adapter.render_feed(&[], &channel("Feed")).unwrap();
+    assert!(xml.contains("<link>http://example.com/users/abc</link>"));
+    assert!(xml.contains(r#"xmlns:atom="http://www.w3.org/2005/Atom""#));
+    assert!(xml.contains(r#"href="http://example.com/users/abc/feed.rss""#));
+    assert!(xml.contains(r#"rel="self""#));
+    assert!(xml.contains("<language>en</language>"));
+}
+
+#[test]
+fn item_links_to_absolute_movie_page() {
+    let adapter = RssAdapter::new("http://example.com".into());
+    let e = entry(false);
+    let xml = adapter.render_feed(&[e.clone()], &channel("Feed")).unwrap();
+    let expected = format!(
+        "<link>http://example.com/movies/{}</link>",
+        e.movie().id().value()
+    );
+    assert!(xml.contains(&expected), "{xml}");
+}
+
+#[test]
+fn spoiler_comment_is_omitted() {
+    let adapter = RssAdapter::new("http://example.com".into());
+    let xml = adapter
+        .render_feed(&[entry(true)], &channel("Feed"))
+        .unwrap();
+    assert!(xml.contains("contains spoilers"));
+    assert!(!xml.contains("replicants"));
+}
diff --git a/crates/domain/src/ports/rss.rs b/crates/domain/src/ports/rss.rs
index 5b420476..3062a934 100644
--- a/crates/domain/src/ports/rss.rs
+++ b/crates/domain/src/ports/rss.rs
@@ -1,5 +1,19 @@
 use crate::models::DiaryEntry;
 
-pub trait RssFeedRenderer: Send + Sync {
-    fn render_feed(&self, entries: &[DiaryEntry], title: &str) -> Result<String, String>;
+/// Channel-level metadata for a rendered feed. Paths are site-relative
+/// (e.g. `/users/{id}`); the renderer resolves them against its base URL.
+pub struct FeedChannel<'a> {
+    pub title: &'a str,
+    /// Human-readable page the channel represents.
+    pub link_path: &'a str,
+    /// Where the feed itself is served, advertised as `atom:link rel="self"`.
+    pub self_path: &'a str,
+}
+
+pub trait RssFeedRenderer: Send + Sync {
+    fn render_feed(
+        &self,
+        entries: &[DiaryEntry],
+        channel: &FeedChannel<'_>,
+    ) -> Result<String, String>;
 }
diff --git a/crates/presentation/src/handlers/rss.rs b/crates/presentation/src/handlers/rss.rs
index d83d7650..ed4d13e4 100644
--- a/crates/presentation/src/handlers/rss.rs
+++ b/crates/presentation/src/handlers/rss.rs
@@ -6,7 +6,7 @@ use axum::{
 use uuid::Uuid;
 
 use application::{diary::get_diary, diary::get_user_feed, diary::queries::GetDiaryQuery};
-use domain::{errors::DomainError, models::ReviewSortBy};
+use domain::{errors::DomainError, models::ReviewSortBy, ports::FeedChannel};
 
 use crate::{errors::ApiError, state::AppState};
 
@@ -21,10 +21,17 @@ pub async fn get_feed(State(state): State<AppState>) -> Result<impl IntoResponse
     let page = get_diary::execute(&state.app_ctx.deps.diary.get_diary, query).await?;
     let xml = state
         .rss_renderer
-        .render_feed(&page.items, "Movie Diary")
+        .render_feed(
+            &page.items,
+            &FeedChannel {
+                title: "Movie Diary",
+                link_path: "/",
+                self_path: "/feed.rss",
+            },
+        )
         .map_err(|e| ApiError(DomainError::InfrastructureError(e)))?;
     Ok((
-        [(header::CONTENT_TYPE, "application/rss+xml; charset=utf-8")],
+        [(header::CONTENT_TYPE, "application/xml; charset=utf-8")],
         xml,
     ))
 }
@@ -41,14 +48,23 @@ pub async fn get_user_feed(
     .await?;
 
     let title = format!("{}'s Movie Diary", feed.author.display_name);
+    let link_path = format!("/users/{user_id}");
+    let self_path = format!("/users/{user_id}/feed.rss");
 
     let xml = state
         .rss_renderer
-        .render_feed(&feed.entries, &title)
+        .render_feed(
+            &feed.entries,
+            &FeedChannel {
+                title: &title,
+                link_path: &link_path,
+                self_path: &self_path,
+            },
+        )
         .map_err(|e| ApiError(DomainError::InfrastructureError(e)))?;
 
     Ok((
-        [(header::CONTENT_TYPE, "application/rss+xml; charset=utf-8")],
+        [(header::CONTENT_TYPE, "application/xml; charset=utf-8")],
         xml,
     ))
 }
diff --git a/crates/presentation/src/tests/extractors.rs b/crates/presentation/src/tests/extractors.rs
index 7ea73b46..8a2cd575 100644
--- a/crates/presentation/src/tests/extractors.rs
+++ b/crates/presentation/src/tests/extractors.rs
@@ -416,7 +416,11 @@ impl domain::ports::DocumentParser for Panic {
 }
 
 impl domain::ports::RssFeedRenderer for Panic {
-    fn render_feed(&self, _: &[DiaryEntry], _: &str) -> Result<String, String> {
+    fn render_feed(
+        &self,
+        _: &[DiaryEntry],
+        _: &domain::ports::FeedChannel<'_>,
+    ) -> Result<String, String> {
         panic!()
     }
 }