diff --git a/docs/forge.md b/docs/forge.md
index 527ca5ac..2b10ce59 100644
--- a/docs/forge.md
+++ b/docs/forge.md
@@ -51,10 +51,16 @@ read it without touching c0re's host-side credential store.
 
 Two things live in the `agent-configs` Forgejo organization:
 
-- A mirror repo per agent (`agent-configs/<name>`) — c0re pushes the
-  agent's applied config repo on each `↻ R3BU1LD`. Agents are
-  read-only collaborators on `core/meta` (the hive-c0re-owned meta
-  flake) so they can fetch but never push.
+- A config repo per agent (`agent-configs/<name>`). As of #1787 the
+  agent is a **write collaborator on its own** repo — it can push
+  config-change branches and (once #1838 P2 lands) open config PRs — but
+  `main` is branch-protected core-only: only hive-c0re's verify-and-ff-push
+  merge handler lands on `main`, an operator-team approval is required, and
+  the agent can neither push `main` directly nor self-merge. During the
+  transition c0re still force-mirrors the agent's applied config repo here
+  on each `↻ R3BU1LD` (the `enable_force_push` allowance), until the PR
+  flow replaces that. Repos stay private, so an agent can't read another
+  agent's config. (Agents remain read-only collaborators on `core/meta`.)
 - The dashboard links each container's "config" anchor to this
   mirror, so operators can click straight from the SW4RM tab into
   the rendered repo without an extra `git` step.
diff --git a/hive-c0re/src/forge.rs b/hive-c0re/src/forge.rs
index e2d953c1..fe330c9b 100644
--- a/hive-c0re/src/forge.rs
+++ b/hive-c0re/src/forge.rs
@@ -30,12 +30,16 @@ const CORE_TOKEN_PATH: &str = "/var/lib/hyperhive/forge-core-token";
 // helpers in `hive_sh4re::assets`. The `agent-configs.png` is
 // rendered from its SVG during the `hyperhive-assets` derivation's
 // build.
-/// Forgejo org grouping every agent's applied config repo. Core is a
-/// site admin and reads + writes every repo here; agents are NOT
-/// members and the repos are private, so no agent — not even the one
-/// a repo describes — can reach a config repo through the forge. The
-/// applied repos stay hive-c0re-owned on disk; this org is just a
-/// mirror target core pushes to.
+/// Forgejo org grouping every agent's config repo. Core is a site admin
+/// and reads + writes every repo here. As of #1787 each agent is a **write
+/// collaborator on its own** `agent-configs/<name>` repo — the editable
+/// PR surface it pushes config-change branches to — but `main` is
+/// branch-protected core-only, so only hive-c0re's verify-and-ff-push merge
+/// handler lands on it (operator approval required; the agent can't push
+/// `main` or self-merge). The repos remain private, so an agent still can't
+/// reach *another* agent's config. During the transition `push_config` also
+/// force-mirrors `applied → main` here until the PR flow (#1838 P2) replaces
+/// it.
 const CONFIG_ORG: &str = "agent-configs";
 /// Forgejo org hosting the operator-curated shared docs/skills repo
 /// that every agent gets read-only access to. Agents use it as a
@@ -688,9 +692,12 @@ pub async fn push_meta(dir: &Path) -> Result<()> {
 }
 
 /// Ensure the `agent-configs/<name>` repo exists so the first
-/// `push_config` doesn't 404. No-op when the forge isn't running or
-/// the core token isn't minted yet. Safe to call on every spawn and
-/// on every startup.
+/// `push_config` doesn't 404, and wire it as the agent-editable PR surface
+/// (#1787): the agent is a **write** collaborator (can push feature branches +
+/// open config PRs) and `main` is branch-protected core-only (only hive-c0re's
+/// merge handler lands on it; operator approval required). No-op when the forge
+/// isn't running or the core token isn't minted yet. Safe to call on every
+/// spawn and on every startup (all steps idempotent).
 pub async fn ensure_config_repo(name: &str) -> Result<()> {
     if !is_present().await {
         return Ok(());
@@ -698,7 +705,12 @@ pub async fn ensure_config_repo(name: &str) -> Result<()> {
     let Some(token) = core_token() else {
         return Ok(());
     };
-    ensure_org_repo(CONFIG_ORG, name, &token).await
+    ensure_org_repo(CONFIG_ORG, name, &token).await?;
+    // Agent = write collaborator: it can push config-PR branches + open PRs,
+    // but the branch protection below keeps it off `main` directly.
+    add_collaborator(CONFIG_ORG, name, name, "write", &token).await?;
+    // Protect `main` core-only (+ transitional force-push for push_config).
+    apply_config_repo_branch_protection(name, &token).await
 }
 
 /// Ensure the `internal/docs` repo exists. Called once at startup
@@ -978,6 +990,45 @@ async fn apply_operator_branch_protection(repo: &str, token: &str) -> Result<()>
     }
 }
 
+/// Apply branch protection to an `agent-configs/<name>` repo's `main` so it
+/// can serve as the agent-editable, PR-merge config surface (#1787 / #1838):
+/// - **push + merge whitelists are `core`-only** — the agent (a write
+///   collaborator) can push feature branches and open config PRs, but only
+///   hive-c0re lands on `main`, via its verify-and-ff-push merge handler
+///   (`run_merge_config_pr`). The agent can never push `main` directly.
+/// - **operator-team approval is required** to merge, and the author (not in
+///   the team) cannot self-approve.
+/// - **`enable_force_push` is a TRANSITIONAL allowance**: `push_config` still
+///   force-mirrors `applied → main` until the PR flow (#1838 P2) replaces it,
+///   and a protected branch otherwise rejects force-push. Forgejo's force-push
+///   allowlist is deploy-keys-only (no per-user list), so this is a plain
+///   `enable_force_push` toggle — only `core` is in the push-whitelist, so
+///   only `core` can force-push anyway. When `push_config` is retired (#1838
+///   P2), flip this to `false`; keep `core` in the push-whitelist so
+///   `ff_push_to_main` can still land fast-forwards.
+///
+/// Idempotent: an existing rule for the branch (200/409/422) is success.
+async fn apply_config_repo_branch_protection(repo: &str, token: &str) -> Result<()> {
+    let url = format!("{FORGE_HTTP}/api/v1/repos/{CONFIG_ORG}/{repo}/branch_protections");
+    let body = format!(
+        r#"{{"branch_name":"main","enable_push_whitelist":true,"push_whitelist_usernames":["core"],"enable_merge_whitelist":true,"merge_whitelist_usernames":["core"],"enable_approvals_whitelist":true,"approvals_whitelist_teams":["{OPERATORS_TEAM}"],"required_approvals":1,"block_on_official_review_requests":true,"allow_manual_merge":true,"enable_force_push":true}}"#
+    );
+    let status = forge_http(reqwest::Method::POST, &url, token, &body).await?;
+    match status.as_u16() {
+        201 => {
+            tracing::info!(%repo, "forge: applied config-repo branch protection");
+            Ok(())
+        }
+        200 | 409 | 422 => {
+            tracing::debug!(%repo, "forge: config-repo branch protection already present");
+            Ok(())
+        }
+        other => {
+            anyhow::bail!("POST {CONFIG_ORG}/{repo}/branch_protections returned HTTP {other}")
+        }
+    }
+}
+
 /// Create a repo for `agent` in the c0re-owned [`AGENTS_ORG`] and wire the
 /// perms: the org owns it (perms stay c0re-managed), the agent is added
 /// as a **write** collaborator (not owner — can push + open PRs but can't