fix: generate agents with platform LLM grants

agent_builder/skills/deepagent-agent-design/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: deepagent-agent-design
-description: Design generated A2A agents around DeepAgents skills, subagents, durable workspace backends, and caller-provided LLM reasoning. Use whenever building or modifying an agent, deciding whether to add tools, creating skill bundles, wiring create_deep_agent, or avoiding shallow fake tools.
+description: Design generated A2A agents around DeepAgents skills, subagents, durable workspace backends, and ctx.llm-backed reasoning. Use whenever building or modifying an agent, deciding whether to add tools, creating skill bundles, wiring create_deep_agent, or avoiding shallow fake tools.
 ---
 # DeepAgent Agent Design
 
@@ -13,8 +13,8 @@ planning, file work, skill selection, and subagent delegation.
 Use this shape for non-trivial generated agents:
 
 1. Keep one or two public A2A `@skill` methods with clear typed parameters.
-2. Inside each method, read `ctx.llm`, construct `ChatOpenAI`, and build a
-   DeepAgent with `create_deep_agent`.
+2. Inside each method, read `ctx.llm`, verify credentials are available before
+   constructing `ChatOpenAI`, and build a DeepAgent with `create_deep_agent`.
 3. Pass `backend=ctx.workspace_backend()` so DeepAgents file tools write to the
    caller's durable workspace instead of LangGraph state.
 4. Seed project skills into the current grant's write prefix and pass