diff --git a/crates/blue-core/src/store.rs b/crates/blue-core/src/store.rs
index 30f1805..23f3246 100644
--- a/crates/blue-core/src/store.rs
+++ b/crates/blue-core/src/store.rs
@@ -124,6 +124,24 @@ const SCHEMA: &str = r#"
 
     CREATE INDEX IF NOT EXISTS idx_staging_locks_resource ON staging_locks(resource);
     CREATE INDEX IF NOT EXISTS idx_staging_queue_resource ON staging_lock_queue(resource);
+
+    CREATE TABLE IF NOT EXISTS staging_deployments (
+        id INTEGER PRIMARY KEY AUTOINCREMENT,
+        name TEXT NOT NULL UNIQUE,
+        iac_type TEXT NOT NULL,
+        deploy_command TEXT NOT NULL,
+        stacks TEXT,
+        deployed_by TEXT NOT NULL,
+        agent_id TEXT,
+        deployed_at TEXT NOT NULL,
+        ttl_expires_at TEXT NOT NULL,
+        status TEXT NOT NULL DEFAULT 'deployed',
+        destroyed_at TEXT,
+        metadata TEXT
+    );
+
+    CREATE INDEX IF NOT EXISTS idx_staging_deployments_status ON staging_deployments(status);
+    CREATE INDEX IF NOT EXISTS idx_staging_deployments_expires ON staging_deployments(ttl_expires_at);
 "#;
 
 /// FTS5 schema for full-text search
@@ -429,6 +447,45 @@ pub enum StagingLockResult {
     },
 }
 
+/// A tracked staging deployment with TTL
+#[derive(Debug, Clone)]
+pub struct StagingDeployment {
+    pub id: Option<i64>,
+    pub name: String,
+    pub iac_type: String,
+    pub deploy_command: String,
+    pub stacks: Option<String>,
+    pub deployed_by: String,
+    pub agent_id: Option<String>,
+    pub deployed_at: String,
+    pub ttl_expires_at: String,
+    pub status: String,
+    pub destroyed_at: Option<String>,
+    pub metadata: Option<String>,
+}
+
+/// Result of staging resource cleanup operation
+#[derive(Debug, Clone)]
+pub struct StagingCleanupResult {
+    /// Number of expired locks deleted
+    pub locks_cleaned: usize,
+    /// Number of orphaned queue entries deleted
+    pub queue_entries_cleaned: usize,
+    /// Number of deployments marked as expired
+    pub deployments_marked_expired: usize,
+    /// Deployments that are expired but not yet destroyed
+    pub expired_deployments_pending_destroy: Vec<ExpiredDeploymentInfo>,
+}
+
+/// Info about an expired deployment that needs to be destroyed
+#[derive(Debug, Clone)]
+pub struct ExpiredDeploymentInfo {
+    pub name: String,
+    pub iac_type: String,
+    pub deploy_command: String,
+    pub stacks: Option<String>,
+}
+
 /// Store errors - in Blue's voice
 #[derive(Debug, thiserror::Error)]
 pub enum StoreError {
@@ -1640,6 +1697,227 @@ impl DocumentStore {
             Ok((locks_cleaned, queue_cleaned))
         })
     }
+
+    // ===== Staging Deployments =====
+
+    /// Record a new staging deployment
+    pub fn record_staging_deployment(
+        &self,
+        name: &str,
+        iac_type: &str,
+        deploy_command: &str,
+        stacks: Option<&str>,
+        deployed_by: &str,
+        agent_id: Option<&str>,
+        ttl_hours: u32,
+        metadata: Option<&str>,
+    ) -> Result<StagingDeployment, StoreError> {
+        self.with_retry(|| {
+            let now = chrono::Utc::now();
+            let ttl_expires = now + chrono::Duration::hours(ttl_hours as i64);
+
+            self.conn.execute(
+                "INSERT OR REPLACE INTO staging_deployments
+                 (name, iac_type, deploy_command, stacks, deployed_by, agent_id, deployed_at, ttl_expires_at, status, metadata)
+                 VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7, ?8, 'deployed', ?9)",
+                params![
+                    name,
+                    iac_type,
+                    deploy_command,
+                    stacks,
+                    deployed_by,
+                    agent_id,
+                    now.to_rfc3339(),
+                    ttl_expires.to_rfc3339(),
+                    metadata
+                ],
+            )?;
+
+            Ok(StagingDeployment {
+                id: Some(self.conn.last_insert_rowid()),
+                name: name.to_string(),
+                iac_type: iac_type.to_string(),
+                deploy_command: deploy_command.to_string(),
+                stacks: stacks.map(|s| s.to_string()),
+                deployed_by: deployed_by.to_string(),
+                agent_id: agent_id.map(|s| s.to_string()),
+                deployed_at: now.to_rfc3339(),
+                ttl_expires_at: ttl_expires.to_rfc3339(),
+                status: "deployed".to_string(),
+                destroyed_at: None,
+                metadata: metadata.map(|s| s.to_string()),
+            })
+        })
+    }
+
+    /// List all staging deployments, optionally filtered by status
+    pub fn list_staging_deployments(
+        &self,
+        status: Option<&str>,
+    ) -> Result<Vec<StagingDeployment>, StoreError> {
+        let query = if status.is_some() {
+            "SELECT id, name, iac_type, deploy_command, stacks, deployed_by, agent_id,
+                    deployed_at, ttl_expires_at, status, destroyed_at, metadata
+             FROM staging_deployments WHERE status = ?1 ORDER BY deployed_at DESC"
+        } else {
+            "SELECT id, name, iac_type, deploy_command, stacks, deployed_by, agent_id,
+                    deployed_at, ttl_expires_at, status, destroyed_at, metadata
+             FROM staging_deployments ORDER BY deployed_at DESC"
+        };
+
+        let mut stmt = self.conn.prepare(query)?;
+
+        let rows = if let Some(s) = status {
+            stmt.query_map(params![s], Self::map_staging_deployment)?
+        } else {
+            stmt.query_map([], Self::map_staging_deployment)?
+        };
+
+        rows.collect::<Result<Vec<_>, _>>()
+            .map_err(StoreError::Database)
+    }
+
+    /// Get expired deployments that need cleanup
+    pub fn get_expired_deployments(&self) -> Result<Vec<StagingDeployment>, StoreError> {
+        let now = chrono::Utc::now().to_rfc3339();
+
+        let mut stmt = self.conn.prepare(
+            "SELECT id, name, iac_type, deploy_command, stacks, deployed_by, agent_id,
+                    deployed_at, ttl_expires_at, status, destroyed_at, metadata
+             FROM staging_deployments
+             WHERE status = 'deployed' AND ttl_expires_at < ?1
+             ORDER BY ttl_expires_at",
+        )?;
+
+        let rows = stmt.query_map(params![now], Self::map_staging_deployment)?;
+
+        rows.collect::<Result<Vec<_>, _>>()
+            .map_err(StoreError::Database)
+    }
+
+    /// Mark a deployment as destroyed
+    pub fn mark_deployment_destroyed(&self, name: &str) -> Result<(), StoreError> {
+        self.with_retry(|| {
+            let now = chrono::Utc::now().to_rfc3339();
+
+            let updated = self.conn.execute(
+                "UPDATE staging_deployments SET status = 'destroyed', destroyed_at = ?1
+                 WHERE name = ?2 AND status = 'deployed'",
+                params![now, name],
+            )?;
+
+            if updated == 0 {
+                return Err(StoreError::NotFound(format!(
+                    "Deployment '{}' not found or already destroyed",
+                    name
+                )));
+            }
+
+            Ok(())
+        })
+    }
+
+    /// Mark expired deployments as expired (for auto-cleanup tracking)
+    pub fn mark_expired_deployments(&self) -> Result<Vec<StagingDeployment>, StoreError> {
+        let expired = self.get_expired_deployments()?;
+
+        self.with_retry(|| {
+            let now = chrono::Utc::now().to_rfc3339();
+            self.conn.execute(
+                "UPDATE staging_deployments SET status = 'expired'
+                 WHERE status = 'deployed' AND ttl_expires_at < ?1",
+                params![now],
+            )?;
+            Ok(())
+        })?;
+
+        Ok(expired)
+    }
+
+    /// Get a specific deployment by name
+    pub fn get_staging_deployment(
+        &self,
+        name: &str,
+    ) -> Result<Option<StagingDeployment>, StoreError> {
+        self.conn
+            .query_row(
+                "SELECT id, name, iac_type, deploy_command, stacks, deployed_by, agent_id,
+                    deployed_at, ttl_expires_at, status, destroyed_at, metadata
+                 FROM staging_deployments WHERE name = ?1",
+                params![name],
+                Self::map_staging_deployment,
+            )
+            .optional()
+            .map_err(StoreError::Database)
+    }
+
+    /// Helper to map a row to StagingDeployment
+    fn map_staging_deployment(row: &rusqlite::Row) -> rusqlite::Result<StagingDeployment> {
+        Ok(StagingDeployment {
+            id: Some(row.get(0)?),
+            name: row.get(1)?,
+            iac_type: row.get(2)?,
+            deploy_command: row.get(3)?,
+            stacks: row.get(4)?,
+            deployed_by: row.get(5)?,
+            agent_id: row.get(6)?,
+            deployed_at: row.get(7)?,
+            ttl_expires_at: row.get(8)?,
+            status: row.get(9)?,
+            destroyed_at: row.get(10)?,
+            metadata: row.get(11)?,
+        })
+    }
+
+    /// Clean up all expired staging resources (locks, deployments, queue entries)
+    pub fn cleanup_expired_staging_resources(&self) -> Result<StagingCleanupResult, StoreError> {
+        let now = chrono::Utc::now().to_rfc3339();
+
+        // Clean up expired locks
+        let locks_cleaned = self.conn.execute(
+            "DELETE FROM staging_locks WHERE expires_at < ?1",
+            params![now],
+        )?;
+
+        // Clean up orphaned queue entries
+        let queue_cleaned = self.conn.execute(
+            "DELETE FROM staging_lock_queue
+             WHERE resource NOT IN (SELECT resource FROM staging_locks)",
+            [],
+        )?;
+
+        // Mark expired deployments
+        let deployments_marked = self.conn.execute(
+            "UPDATE staging_deployments SET status = 'expired'
+             WHERE status = 'deployed' AND ttl_expires_at < ?1",
+            params![now],
+        )?;
+
+        // Get list of expired deployments that need cleanup commands
+        let mut stmt = self.conn.prepare(
+            "SELECT name, iac_type, deploy_command, stacks
+             FROM staging_deployments
+             WHERE status = 'expired' AND destroyed_at IS NULL",
+        )?;
+
+        let expired_deployments: Vec<ExpiredDeploymentInfo> = stmt
+            .query_map([], |row| {
+                Ok(ExpiredDeploymentInfo {
+                    name: row.get(0)?,
+                    iac_type: row.get(1)?,
+                    deploy_command: row.get(2)?,
+                    stacks: row.get(3)?,
+                })
+            })?
+            .collect::<Result<Vec<_>, _>>()?;
+
+        Ok(StagingCleanupResult {
+            locks_cleaned,
+            queue_entries_cleaned: queue_cleaned,
+            deployments_marked_expired: deployments_marked,
+            expired_deployments_pending_destroy: expired_deployments,
+        })
+    }
 }
 
 #[cfg(test)]
diff --git a/crates/blue-mcp/src/handlers/staging.rs b/crates/blue-mcp/src/handlers/staging.rs
index da198da..e9098c6 100644
--- a/crates/blue-mcp/src/handlers/staging.rs
+++ b/crates/blue-mcp/src/handlers/staging.rs
@@ -359,6 +359,79 @@ pub fn handle_cost(args: &Value, repo_path: &std::path::Path) -> Result<Value, S
     }))
 }
 
+/// Handle blue_staging_deployments
+///
+/// Lists staging deployments with optional filtering.
+/// Can also check for expired deployments that need cleanup.
+pub fn handle_deployments(state: &ProjectState, args: &Value) -> Result<Value, ServerError> {
+    let status = args.get("status").and_then(|v| v.as_str());
+    let check_expired = args
+        .get("check_expired")
+        .and_then(|v| v.as_bool())
+        .unwrap_or(false);
+
+    // Check for expired deployments if requested
+    let expired = if check_expired {
+        state
+            .store
+            .mark_expired_deployments()
+            .map_err(|e| ServerError::CommandFailed(e.to_string()))?
+    } else {
+        vec![]
+    };
+
+    // List deployments with optional status filter
+    let deployments = state
+        .store
+        .list_staging_deployments(status)
+        .map_err(|e| ServerError::CommandFailed(e.to_string()))?;
+
+    let deployed_count = deployments
+        .iter()
+        .filter(|d| d.status == "deployed")
+        .count();
+    let expired_count = expired.len();
+
+    let hint = if check_expired && expired_count > 0 {
+        format!(
+            "{} staging deployment(s). {} newly expired and marked for cleanup.",
+            deployments.len(),
+            expired_count
+        )
+    } else {
+        format!(
+            "{} staging deployment(s) ({} active).",
+            deployments.len(),
+            deployed_count
+        )
+    };
+
+    Ok(json!({
+        "status": "success",
+        "message": blue_core::voice::info(
+            &format!("{} staging deployment{}", deployments.len(), if deployments.len() == 1 { "" } else { "s" }),
+            Some(&hint)
+        ),
+        "deployments": deployments.iter().map(|d| json!({
+            "name": d.name,
+            "iac_type": d.iac_type,
+            "deploy_command": d.deploy_command,
+            "stacks": d.stacks,
+            "deployed_by": d.deployed_by,
+            "deployed_at": d.deployed_at,
+            "ttl_expires_at": d.ttl_expires_at,
+            "status": d.status,
+            "destroyed_at": d.destroyed_at,
+        })).collect::<Vec<_>>(),
+        "expired_count": expired_count,
+        "newly_expired": expired.iter().map(|d| json!({
+            "name": d.name,
+            "iac_type": d.iac_type,
+            "ttl_expires_at": d.ttl_expires_at,
+        })).collect::<Vec<_>>(),
+    }))
+}
+
 fn detect_cdk_stacks(path: &std::path::Path) -> Vec<String> {
     let mut stacks = Vec::new();
 
diff --git a/crates/blue-mcp/src/server.rs b/crates/blue-mcp/src/server.rs
index e0b6b08..8e03b84 100644
--- a/crates/blue-mcp/src/server.rs
+++ b/crates/blue-mcp/src/server.rs
@@ -849,6 +849,28 @@ impl BlueServer {
                         }
                     }
                 },
+                {
+                    "name": "blue_staging_deployments",
+                    "description": "List staging environment deployments. Shows deployed, destroyed, or expired environments. Use check_expired=true to mark expired deployments.",
+                    "inputSchema": {
+                        "type": "object",
+                        "properties": {
+                            "cwd": {
+                                "type": "string",
+                                "description": "Current working directory"
+                            },
+                            "status": {
+                                "type": "string",
+                                "enum": ["deployed", "destroyed", "expired"],
+                                "description": "Filter by deployment status"
+                            },
+                            "check_expired": {
+                                "type": "boolean",
+                                "description": "Check for and mark expired deployments (default: false)"
+                            }
+                        }
+                    }
+                },
                 {
                     "name": "blue_audit",
                     "description": "Check project health and find issues. Returns stalled work, missing ADRs, and recommendations.",
@@ -1387,6 +1409,7 @@ impl BlueServer {
             "blue_staging_unlock" => self.handle_staging_unlock(&call.arguments),
             "blue_staging_status" => self.handle_staging_status(&call.arguments),
             "blue_staging_cleanup" => self.handle_staging_cleanup(&call.arguments),
+            "blue_staging_deployments" => self.handle_staging_deployments(&call.arguments),
             // Phase 6: Audit and completion handlers
             "blue_audit" => self.handle_audit(&call.arguments),
             "blue_rfc_complete" => self.handle_rfc_complete(&call.arguments),
@@ -1955,6 +1978,13 @@ impl BlueServer {
         crate::handlers::staging::handle_cleanup(state, args)
     }
 
+    fn handle_staging_deployments(&mut self, args: &Option<Value>) -> Result<Value, ServerError> {
+        let empty = json!({});
+        let args = args.as_ref().unwrap_or(&empty);
+        let state = self.ensure_state()?;
+        crate::handlers::staging::handle_deployments(state, args)
+    }
+
     // Phase 6: Audit and completion handlers
 
     fn handle_audit(&mut self, _args: &Option<Value>) -> Result<Value, ServerError> {