diff --git a/commands/gsd/query-intel.md b/commands/gsd/query-intel.md
new file mode 100644
index 000000000..e25b12c0f
--- /dev/null
+++ b/commands/gsd/query-intel.md
@@ -0,0 +1,128 @@
+---
+name: gsd:query-intel
+description: Query codebase intelligence graph for dependencies and hotspots
+argument-hint: "<dependents|hotspots> [file-path]"
+allowed-tools:
+  - Bash
+---
+
+<objective>
+Query the codebase intelligence graph database for relationship information.
+
+**Query types:**
+- `dependents <file>` — What files depend on this file? (blast radius)
+- `hotspots` — Which files have the most dependents? (change carefully)
+
+Output: Formatted query results from graph.db
+</objective>
+
+<context>
+This command exposes the graph query capabilities built by Phase 4 (Semantic Intelligence).
+
+**Use for:**
+- Checking blast radius before refactoring a core file
+- Identifying high-impact files that need careful changes
+- Understanding dependency relationships in the codebase
+
+**Requires:** `.planning/intel/graph.db` (created by `/gsd:analyze-codebase` with entity generation)
+
+If graph.db doesn't exist, the command will return an error suggesting to run analyze-codebase first.
+</context>
+
+<process>
+
+## Step 1: Parse arguments
+
+Extract query type and optional file path from arguments.
+
+**Arguments:** $ARGUMENTS
+
+**Expected formats:**
+- `dependents src/lib/db.ts` — query what depends on this file
+- `hotspots` — query most-depended-on files
+- `hotspots 10` — query top 10 hotspots (default: 5)
+
+## Step 2: Convert file path to entity ID
+
+For `dependents` queries, convert the file path to entity ID format:
+- `src/lib/db.ts` → `src-lib-db`
+- Replace `/` with `-`, remove extension
+
+```bash
+# Example conversion
+FILE_PATH="src/lib/db.ts"
+ENTITY_ID=$(echo "$FILE_PATH" | sed 's/\.[^.]*$//' | tr '/' '-')
+```
+
+## Step 3: Execute query
+
+Run the appropriate query against the graph database:
+
+**For dependents:**
+```bash
+echo '{"action":"query","type":"dependents","target":"'$ENTITY_ID'","limit":20}' | node hooks/gsd-intel-index.js
+```
+
+**For hotspots:**
+```bash
+echo '{"action":"query","type":"hotspots","limit":5}' | node hooks/gsd-intel-index.js
+```
+
+## Step 4: Format and present results
+
+Parse the JSON response and present in readable format.
+
+**For dependents:**
+```
+## Files that depend on {file-path}
+
+Found {count} dependents:
+
+1. src/api/users.ts
+2. src/api/auth.ts
+3. src/services/payment.ts
+...
+
+**Blast radius:** {count} files would be affected by changes.
+```
+
+**For hotspots:**
+```
+## Dependency Hotspots
+
+These files have the most dependents — change carefully:
+
+| Rank | File | Dependents |
+|------|------|------------|
+| 1 | src/lib/db.ts | 42 |
+| 2 | src/types/user.ts | 35 |
+| 3 | src/utils/format.ts | 28 |
+```
+
+## Step 5: Handle errors
+
+**If graph.db doesn't exist:**
+```
+No graph database found at .planning/intel/graph.db
+
+Run /gsd:analyze-codebase first to build the dependency graph.
+```
+
+**If entity not found:**
+```
+No entity found for: {file-path}
+
+The file may not be indexed yet. Try:
+- /gsd:analyze-codebase to rebuild the index
+- Check the file path is correct
+```
+
+</process>
+
+<success_criteria>
+- [ ] Query type parsed from arguments
+- [ ] File path converted to entity ID (for dependents)
+- [ ] Query executed against graph.db
+- [ ] Results formatted in readable markdown
+- [ ] Errors handled gracefully with helpful messages
+</success_criteria>