docs: document the new compression_model setting (global + per-agent override)

Agents.md

@@ -61,6 +61,7 @@ top_p: null                          # Set default top-p parameter, with a range
 reasoning_effort: null               # Set the reasoning effort level for models that support it (e.g. low, medium, high)
 # Agent Metadata Configuration
 agent_session: null                  # Set a session to use when starting the agent. (e.g. temp, default); defaults to globally set agent_session
+compression_model: null              # Model to use when compressing this agent's session (overrides the global compression_model); falls back to the current model with a printed warning if it fails
 # Agent Configuration
 name: <agent-name>                   # Name of the agent, used in the UI and logs
 description: <description>           # Description of the agent, used in the UI

Sessions.md

@@ -116,6 +116,7 @@ The following settings are available to customize the default behavior of sessio
 | `summarization_prompt`   | This is the prompt that is used to compress the session up to a given point when compression is triggered                                                                                                                                                                                                                                                              |
 | `summary_context_prompt` | This is the prompt that's used to add the summarized conversation generated by the `summarization_prompt` as context to the model                                                                                                                                                                                                                                      |
 | `compression_keep_last`  | Number of the most recent messages to keep visible in the session after compression is triggered. Earlier messages are folded into the summary. `0` (the default) compresses all messages.                                                                                                                                                                             |
+| `compression_model`      | Model to use for the compression (summarization) LLM call (e.g. `openai:gpt-4o-mini`), letting you use a cheaper/faster model than the session's current model. If the specified model fails (unknown model, API error, or timeout), Coyote prints a warning and falls back to the session's current model. `null` (the default) uses the session's current model. Agents can override this per-agent via `compression_model` in their `config.yaml`.       |
 | `max_tool_result_chars`  | Caps the number of characters from any single tool result forwarded to the model. Results that exceed this limit are truncated and a notice is prepended. Useful for preventing large tool outputs from consuming context. `null` (the default) disables the cap.                                                                                                       |
 
 ## Per-Session Settings