Chore/codex max release review

README.md

@@ -450,18 +450,18 @@ When using [`config/full-opencode.json`](./config/full-opencode.json), you get t
 
 | CLI Model ID | TUI Display Name | Reasoning Effort | Best For |
 |--------------|------------------|-----------------|----------|
-| `gpt-5.1-codex-max` | GPT 5.1 Codex Max (OAuth) | Medium (Extra High optional) | Default flagship tier with optional `xhigh` reasoning for long, complex runs |
+| `gpt-5.1-codex-max` | GPT 5.1 Codex Max (OAuth) | Low/Medium/High/**Extra High** | Default flagship tier with `xhigh` reasoning for complex, multi-step problems |
 | `gpt-5.1-codex-low` | GPT 5.1 Codex Low (OAuth) | Low | Fast code generation on the newest Codex tier |
 | `gpt-5.1-codex-medium` | GPT 5.1 Codex Medium (OAuth) | Medium | Balanced code + tooling workflows |
 | `gpt-5.1-codex-high` | GPT 5.1 Codex High (OAuth) | High | Multi-step coding tasks with deep tool use |
 | `gpt-5.1-codex-mini-medium` | GPT 5.1 Codex Mini Medium (OAuth) | Medium | Budget-friendly Codex runs (200k/100k tokens) |
 | `gpt-5.1-codex-mini-high` | GPT 5.1 Codex Mini High (OAuth) | High | Cheaper Codex tier with maximum reasoning |
-| `gpt-5.1-none` | GPT 5.1 None (OAuth) | None | Latency-sensitive chat/tasks using the new "no reasoning" mode |
+| `gpt-5.1-none` | GPT 5.1 None (OAuth) | **None** | Latency-sensitive chat/tasks using the "no reasoning" mode |
 | `gpt-5.1-low` | GPT 5.1 Low (OAuth) | Low | Fast general-purpose chat with light reasoning |
 | `gpt-5.1-medium` | GPT 5.1 Medium (OAuth) | Medium | Default adaptive reasoning for everyday work |
 | `gpt-5.1-high` | GPT 5.1 High (OAuth) | High | Deep analysis when reliability matters most |
 
-> **Extra High reasoning:** `reasoningEffort: "xhigh"` is exclusive to `gpt-5.1-codex-max`. Other models automatically map that option to `high` so their API calls remain valid.
+> **Extra High reasoning:** `reasoningEffort: "xhigh"` provides maximum computational effort for complex, multi-step problems and is exclusive to `gpt-5.1-codex-max`. Other models automatically map that option to `high` so their API calls remain valid.
 
 #### Legacy GPT-5 lineup (still supported)
 
@@ -540,12 +540,14 @@ If you want to customize settings yourself, you can configure options at provide
 
 | Setting | GPT-5 / GPT-5.1 Values | GPT-5-Codex / Codex Mini Values | Plugin Default |
 |---------|-------------|-------------------|----------------|
-| `reasoningEffort` | `none`, `minimal`, `low`, `medium`, `high` | `low`, `medium`, `high`, `xhigh`* | `medium` |
+| `reasoningEffort` | `none`, `minimal`, `low`, `medium`, `high` | `low`, `medium`, `high`, `xhigh`† | `medium` |
 | `reasoningSummary` | `auto`, `detailed` | `auto`, `detailed` | `auto` |
 | `textVerbosity` | `low`, `medium`, `high` | `medium` only | `medium` |
 | `include` | Array of strings | Array of strings | `["reasoning.encrypted_content"]` |
 
 > **Note**: `minimal` effort is auto-normalized to `low` for gpt-5-codex (not supported by the API). `none` is only supported on GPT-5.1 general models; when used with legacy gpt-5 it is normalized to `minimal`. `xhigh` is exclusive to `gpt-5.1-codex-max`—other Codex presets automatically map it to `high`.
+> 
+> † **Extra High reasoning**: `reasoningEffort: "xhigh"` provides maximum computational effort for complex, multi-step problems and is only available on `gpt-5.1-codex-max`.
 
 #### Plugin-Level Settings
 

docs/development/CONFIG_FIELDS.md

@@ -305,8 +305,8 @@ const parsedModel: ModelsDev.Model = {
 
 **Why this matters:**
 - Config keys mirror the Codex CLI's 5.1 presets, making it obvious which tier you're targeting.
-- `reasoningEffort: "none"` is only valid for GPT-5.1 general models—the plugin automatically downgrades unsupported values for Codex/Codex Mini.
-- `reasoningEffort: "xhigh"` is exclusive to `gpt-5.1-codex-max`; other models automatically clamp it to `high`.
+- `reasoningEffort: "none"` (No Reasoning) disables reasoning entirely for latency-sensitive tasks and is only valid for GPT-5.1 general models—the plugin automatically downgrades unsupported values for Codex/Codex Mini.
+- `reasoningEffort: "xhigh"` (Extra High) provides maximum computational effort for complex, multi-step problems and is exclusive to `gpt-5.1-codex-max`; other models automatically clamp it to `high`.
 - Legacy GPT-5, GPT-5-Codex, and Codex Mini presets automatically clamp unsupported values (`none` → `minimal`/`low`, `minimal` → `low` for Codex).
 - Mixing GPT-5.1 and GPT-5 presets inside the same config is fine—just keep config keys unique and let the plugin normalize them.
 

docs/notes/2025.11.19.18.38.24.md

@@ -0,0 +1,86 @@
+In lib/request/request-transformer.ts around lines 952 to 968, the
+TransformResult interface is currently unexported which prevents consumers from
+using the return type of transformRequestBody; change the declaration to export
+interface TransformResult so it’s exported from the module, and update any local
+references or imports elsewhere if needed to use the exported type (no other
+logic changes).
+
+
+In lib/request/request-transformer.ts around lines 621 to 633, the code
+duplicates the bridge message object creation (the developer role message with
+CODEX_OPENCODE_BRIDGE and input merging) which is repeated later at lines
+~658-667; extract a small helper (e.g., buildBridgeMessage(input):
+Array<Message> or getBridgeMessage(input): Message[]) that returns the array
+with the bridge message followed by the existing input and replace both
+duplicated branches with a call to that helper, keeping existing types and
+imports and ensuring generateContentHash("add") checks still control whether to
+return the helper result or the original input.
+
+
+In lib/compaction/compaction-executor.ts around lines 24 to 66, wrap the
+response.text() + JSON.parse(...) and subsequent payload manipulation in a
+try/catch so non‑JSON or unexpected response shapes do not crash compaction; on
+any parse or processing error, log or ignore the error and return the original
+response object untouched. Ensure the catch block returns the original Response
+(preserving status, statusText, headers, and body) so callers receive the
+unmodified response when parsing fails.
+
+
+In lib/compaction/codex-compaction.ts around lines 168 to 170, the cloneRange
+function duplicates logic already implemented in lib/utils/clone.ts as
+cloneInputItems; replace the local implementation by importing cloneInputItems
+from 'lib/utils/clone' and call it where cloneRange is used (or rename uses to
+cloneInputItems), remove the duplicate function, and ensure the import is added
+and TypeScript types align with InputItem[].
+
+
+In lib/compaction/codex-compaction.ts around lines 131 to 144, the
+extractTextFromItem function duplicates logic already in
+lib/utils/input-item-utils.ts; replace this local implementation by importing
+and calling the centralized utility (ensuring the import path is correct), and
+if needed adapt or wrap the utility call so behavior remains identical (handle
+null/undefined input and array/object type checks the same way as the previous
+local function). Remove the duplicated function, run type checks/TS compile and
+unit tests to confirm no behavioral regressions.
+
+
+lib/cache/cache-metrics.ts lines 34-53 (also apply similar changes at 59-79,
+103-105, 167-185): the metrics object and API are tightened to prevent
+accidental writes to the aggregate bucket but getMetrics currently performs only
+a shallow clone so callers can still mutate nested CacheMetrics; update the
+types to use keyof Omit<CacheMetricsCollection, "overall"> for per-key
+operations (hits/misses/evictions) and ensure every place that updates rates
+also recomputes and updates the "overall" hitRate consistently, and either
+return a deep-cloned/read-only snapshot from getMetrics or clearly document the
+return as read-only to prevent external mutation.
+
+In lib/cache/cache-warming.ts around lines 113 to 126, the catch block declares
+an unused named parameter (_error) causing lint/typecheck warnings; remove the
+unused binding by changing the catch to a bare catch (i.e., catch { ... }) so
+the error is still ignored and the function behavior remains identical while
+satisfying the linter.
+
+In lib/compaction/codex-compaction.ts around lines 131 to 144, the
+extractTextFromItem function duplicates logic already in
+lib/utils/input-item-utils.ts; replace this local implementation by importing
+and calling the centralized utility (ensuring the import path is correct), and
+if needed adapt or wrap the utility call so behavior remains identical (handle
+null/undefined input and array/object type checks the same way as the previous
+local function). Remove the duplicated function, run type checks/TS compile and
+unit tests to confirm no behavioral regressions.
+
+In lib/compaction/codex-compaction.ts around lines 168 to 170, the cloneRange
+function duplicates logic already implemented in lib/utils/clone.ts as
+cloneInputItems; replace the local implementation by importing cloneInputItems
+from 'lib/utils/clone' and call it where cloneRange is used (or rename uses to
+cloneInputItems), remove the duplicate function, and ensure the import is added
+and TypeScript types align with InputItem[].
+
+In lib/compaction/compaction-executor.ts around lines 24 to 66, wrap the
+response.text() + JSON.parse(...) and subsequent payload manipulation in a
+try/catch so non‑JSON or unexpected response shapes do not crash compaction; on
+any parse or processing error, log or ignore the error and return the original
+response object untouched. Ensure the catch block returns the original Response
+(preserving status, statusText, headers, and body) so callers receive the
+unmodified response when parsing fails.
+

docs/reasoning-effort-levels-update.md

@@ -0,0 +1,84 @@
+# Reasoning Effort Levels Documentation Update
+
+## Summary
+
+Update documentation to clearly explain all available reasoning effort levels for the new Codex Max model, including `none`, `low`, `medium`, `high`, and `xhigh`.
+
+## Current State
+
+Based on codebase analysis:
+
+### Already Implemented ✅
+- `xhigh` reasoning effort is supported in code (`lib/types.ts:53`, `lib/request/request-transformer.ts:327`)
+- Tests cover `xhigh` handling (`test/request-transformer.test.ts:141-162`)
+- README.md mentions `xhigh` for Codex Max (`README.md:453,464,543,548`)
+- Configuration files include proper reasoning levels
+- AGENTS.md documents `xhigh` exclusivity to Codex Max
+
+### Documentation Gaps Identified
+1. README.md could be clearer about the complete range of reasoning levels
+2. Need to ensure all reasoning levels (`none`, `low`, `medium`, `high`, `xhigh`) are clearly documented
+3. Configuration examples should show the full spectrum
+
+## Files to Update
+
+### Primary Documentation
+- `README.md` - Main user-facing documentation
+- `docs/development/CONFIG_FIELDS.md` - Developer configuration reference
+
+### Configuration Examples (Already Up-to-Date)
+- `config/full-opencode.json` - Complete configuration with all reasoning levels
+- `config/minimal-opencode.json` - Minimal configuration
+
+## Definition of Done
+
+- [x] All reasoning effort levels (`none`, `low`, `medium`, `high`, `xhigh`) are clearly documented
+- [x] `xhigh` exclusivity to `gpt-5.1-codex-max` is clearly explained
+- [x] Automatic downgrade behavior for unsupported models is documented
+- [x] Configuration examples show the complete range of reasoning levels
+- [x] Documentation is consistent across all files
+
+## Implementation Notes
+
+### Reasoning Effort Levels by Model Type
+
+| Model Type | Supported Levels | Notes |
+|------------|----------------|-------|
+| `gpt-5.1-codex-max` | `low`, `medium`, `high`, `xhigh` | `xhigh` is exclusive to this model |
+| `gpt-5.1-codex` | `low`, `medium`, `high` | `xhigh` auto-downgrades to `high` |
+| `gpt-5.1-codex-mini` | `low`, `medium`, `high` | `xhigh` auto-downgrades to `high` |
+| `gpt-5.1` (general) | `none`, `low`, `medium`, `high` | `none` only supported on general models |
+| `gpt-5-codex` | `low`, `medium`, `high` | `minimal` auto-normalizes to `low` |
+| `gpt-5` (legacy) | `minimal`, `low`, `medium`, `high` | `none` auto-normalizes to `minimal` |
+
+### Automatic Normalization Rules
+
+1. **`xhigh` handling**: Only allowed on `gpt-5.1-codex-max`, others downgrade to `high`
+2. **`none` handling**: Only supported on GPT-5.1 general models, legacy gpt-5 normalizes to `minimal`
+3. **`minimal` handling**: Normalizes to `low` for Codex models (not supported by API)
+
+## Changes Made
+
+### README.md Updates
+- Enhanced reasoning effort documentation table
+- Added clearer explanation of `xhigh` exclusivity
+- Updated model variant descriptions to include reasoning level ranges
+- Improved configuration examples section
+
+### CONFIG_FIELDS.md Updates
+- Added `xhigh` to the reasoning effort documentation
+- Clarified which models support which levels
+- Documented automatic normalization behavior
+
+## Testing Verification
+
+All reasoning effort levels are already tested in:
+- `test/request-transformer.test.ts:141-162` - `xhigh` handling tests
+- `test/request-transformer.test.ts:125-153` - Basic reasoning config tests
+- Integration tests cover full configuration flow
+
+## Impact
+
+- **Users**: Clearer understanding of available reasoning levels and model capabilities
+- **Developers**: Better documentation for configuration options
+- **Support**: Reduced confusion about reasoning effort limitations per model

eslint.config.mjs

@@ -52,7 +52,7 @@ export default [
       "sonarjs/cognitive-complexity": ["error", 30],
 
       // Function and file size limits (line counts ignore blank lines and comments)
-      "max-lines-per-function": ["warn", { max: 80, skipBlankLines: true, skipComments: true }],
+      "max-lines-per-function": ["warn", { max: 120, skipBlankLines: true, skipComments: true }],
       "max-lines": ["warn", { max: 500, skipBlankLines: true, skipComments: true }],
 
       // Rely on TypeScript for undefined/global checks

lib/auth/auth.ts

@@ -109,7 +109,9 @@ export function decodeJWT(token: string): JWTPayload | null {
 		const parts = token.split(".");
 		if (parts.length !== 3) return null;
 		const payload = parts[1];
-		const decoded = Buffer.from(payload, "base64").toString("utf-8");
+		const normalized = payload.replace(/-/g, "+").replace(/_/g, "/");
+		const padded = normalized.padEnd(Math.ceil(normalized.length / 4) * 4, "=");
+		const decoded = Buffer.from(padded, "base64").toString("utf-8");
 		return JSON.parse(decoded) as JWTPayload;
 	} catch {
 		return null;

lib/auth/browser.ts

@@ -29,7 +29,7 @@ export function openBrowserUrl(url: string): void {
 			stdio: "ignore",
 			shell: process.platform === "win32",
 		});
-	} catch (_error) {
+	} catch {
 		// Silently fail - user can manually open the URL from instructions
 	}
 }

lib/auth/server.ts

@@ -16,31 +16,34 @@ const successHtml = fs.readFileSync(path.join(__dirname, "..", "oauth-success.ht
  */
 export function startLocalOAuthServer({ state }: { state: string }): Promise<OAuthServerInfo> {
 	const server = http.createServer((req, res) => {
+		const send = (status: number, message: string, headers?: http.OutgoingHttpHeaders) => {
+			const finalHeaders = {
+				"Content-Type": "text/plain; charset=utf-8",
+				...headers,
+			};
+			res.writeHead(status, finalHeaders);
+			res.end(message);
+		};
+
 		try {
 			const url = new URL(req.url || "", "http://localhost");
 			if (url.pathname !== "/auth/callback") {
-				res.statusCode = 404;
-				res.end("Not found");
+				send(404, "Not found");
 				return;
 			}
 			if (url.searchParams.get("state") !== state) {
-				res.statusCode = 400;
-				res.end("State mismatch");
+				send(400, "State mismatch");
 				return;
 			}
 			const code = url.searchParams.get("code");
 			if (!code) {
-				res.statusCode = 400;
-				res.end("Missing authorization code");
+				send(400, "Missing authorization code");
 				return;
 			}
-			res.statusCode = 200;
-			res.setHeader("Content-Type", "text/html; charset=utf-8");
-			res.end(successHtml);
+			send(200, successHtml, { "Content-Type": "text/html; charset=utf-8" });
 			(server as http.Server & { _lastCode?: string })._lastCode = code;
 		} catch {
-			res.statusCode = 500;
-			res.end("Internal error");
+			send(500, "Internal error");
 		}
 	});
 

lib/cache/cache-metrics.ts

@@ -96,12 +96,22 @@ class CacheMetricsCollector {
 		metrics.hitRate = metrics.totalRequests > 0 ? (metrics.hits / metrics.totalRequests) * 100 : 0;
 	}
 
+	private cloneMetrics(): CacheMetricsCollection {
+		const cloneMetric = (metric: CacheMetrics): CacheMetrics => ({ ...metric });
+		return {
+			codexInstructions: cloneMetric(this.metrics.codexInstructions),
+			opencodePrompt: cloneMetric(this.metrics.opencodePrompt),
+			bridgeDecisions: cloneMetric(this.metrics.bridgeDecisions),
+			overall: cloneMetric(this.metrics.overall),
+		};
+	}
+
 	/**
 	 * Get current metrics
 	 * @returns Complete metrics collection
 	 */
 	getMetrics(): CacheMetricsCollection {
-		return { ...this.metrics };
+		return this.cloneMetrics();
 	}
 
 	/**

lib/cache/cache-warming.ts

@@ -119,7 +119,7 @@ export async function areCachesWarm(): Promise<boolean> {
 
 		// If both caches have valid entries, they are warm
 		return !!(codexEntry && opencodeEntry);
-	} catch (_error) {
+	} catch {
 		// Any error suggests caches are not warm
 		return false;
 	}

lib/cache/session-cache.ts

@@ -65,9 +65,7 @@ export function createSessionCache<T>(ttlMs = 15 * 60 * 1000): SessionCache<T> {
 		}
 	};
 
-	const getSize = (): number => {
-		return cache.size;
-	};
+	const getSize = (): number => cache.size;
 
 	return { get, set, clear, clean, getSize };
 }