Docs: fix internal links

docs/concepts/agent.md

@@ -13,11 +13,11 @@ Clawdbot uses a single agent workspace directory (`agents.defaults.workspace`) a
 
 Recommended: use `clawdbot setup` to create `~/.clawdbot/clawdbot.json` if missing and initialize the workspace files.
 
-Full workspace layout + backup guide: [`docs/agent-workspace.md`](/concepts/agent-workspace)
+Full workspace layout + backup guide: [Agent workspace](/concepts/agent-workspace)
 
 If `agents.defaults.sandbox` is enabled, non-main sessions can override this with
 per-session workspaces under `agents.defaults.sandbox.workspaceRoot` (see
-[`docs/configuration.md`](/gateway/configuration)).
+[Gateway configuration](/gateway/configuration)).
 
 ## Bootstrap files (injected)
 
@@ -54,7 +54,7 @@ Clawdbot loads skills from three locations (workspace wins on name conflict):
 - Managed/local: `~/.clawdbot/skills`
 - Workspace: `<workspace>/skills`
 
-Skills can be gated by config/env (see `skills` in [`docs/configuration.md`](/gateway/configuration)).
+Skills can be gated by config/env (see `skills` in [Gateway configuration](/gateway/configuration)).
 
 ## p-mono integration
 
@@ -81,7 +81,7 @@ message is injected before the next assistant response.
 
 When queue mode is `followup` or `collect`, inbound messages are held until the
 current turn ends, then a new agent turn starts with the queued payloads. See
-[`docs/queue.md`](/concepts/queue) for mode + debounce/cap behavior.
+[Queue](/concepts/queue) for mode + debounce/cap behavior.
 
 Block streaming sends completed assistant blocks as soon as they finish; it is
 **off by default** (`agents.defaults.blockStreamingDefault: "off"`).