[pull] main from regenrek:main

.cursor/commands/ask-gpt-pro-api.md

@@ -0,0 +1,40 @@
+---
+description: Ask GPT‑5 Pro (API) with Codefetch context
+argument-hint: PROMPT="<your question or task>"
+---
+
+# Ask GPT‑5 Pro (API engine)
+
+## Instructions
+
+1. **Research first**: Analyze the PROMPT to identify which files/directories are relevant to the task.
+2. **Determine scope**: Based on your research, decide which files or directories to include:
+   - Use `--include-files` for specific file patterns (supports globs, comma-separated): `"src/auth/**/*.ts,src/api/**/*.ts"`
+   - Use `--include-dir` for entire directories (comma-separated): `"src/auth,src/utils"`
+   - You can combine both options
+3. **Generate context**: Run codefetch with the determined scope, limited to 50k tokens.
+4. **Ask Oracle**: Use the generated markdown file with Oracle.
+
+## Run
+
+```bash
+# 1) Research the codebase to determine relevant files/directories based on PROMPT
+# Then generate context (limit ~50k tokens) using appropriate codefetch options:
+#   --include-files "pattern1,pattern2"  for file globs
+#   --include-dir "dir1,dir2"            for directories
+#   Combine both if needed
+npx codefetch --max-tokens 50000 \
+  --include-files "src/auth/**/*,src/api/**/*" \
+  --include-dir "src/utils"
+
+# 2) Ask Oracle via API (requires OPENAI_API_KEY)
+OPENAI_API_KEY=sk-... npx -y @steipete/oracle --engine api -m gpt-5-pro \
+  -p "$PROMPT" \
+  --file codefetch/codebase.md
+
+# Optional helpers:
+#   --files-report       # print per-file token usage
+#   --preview            # inspect the bundle before sending
+```
+
+

.cursor/commands/ask-gpt-pro-browser.md

@@ -0,0 +1,40 @@
+---
+description: Ask GPT‑5 Pro (Browser) with Codefetch context
+argument-hint: PROMPT="<your question or task>"
+---
+
+# Ask GPT‑5 Pro (Browser engine)
+
+## Instructions
+
+1. **Research first**: Analyze the PROMPT to identify which files/directories are relevant to the task.
+2. **Determine scope**: Based on your research, decide which files or directories to include:
+   - Use `--include-files` for specific file patterns (supports globs, comma-separated): `"src/auth/**/*.ts,src/api/**/*.ts"`
+   - Use `--include-dir` for entire directories (comma-separated): `"src/auth,src/utils"`
+   - You can combine both options
+3. **Generate context**: Run codefetch with the determined scope, limited to 50k tokens.
+4. **Ask Oracle**: Use the generated markdown file with Oracle.
+
+## Run
+
+```bash
+# 1) Research the codebase to determine relevant files/directories based on PROMPT
+# Then generate context (limit ~50k tokens) using appropriate codefetch options:
+#   --include-files "pattern1,pattern2"  for file globs
+#   --include-dir "dir1,dir2"            for directories
+#   Combine both if needed
+npx codefetch --max-tokens 50000 \
+  --include-files "src/auth/**/*,src/api/**/*" \
+  --include-dir "src/utils"
+
+# 2) Ask Oracle via Browser (no API key; macOS + Chrome)
+npx -y @steipete/oracle --engine browser -m gpt-5-pro \
+  -p "$PROMPT" \
+  --file codefetch/codebase.md
+
+# Optional helpers:
+#   --files-report       # print per-file token usage
+#   --preview            # inspect the bundle before sending
+```
+
+

packages/cli/package.json

@@ -40,7 +40,7 @@
   },
   "dependencies": {
     "c12": "^2.0.1",
-    "codefetch-sdk": "^2.0.0",
+    "codefetch-sdk": "workspace:*",
     "consola": "^3.3.3",
     "ignore": "^7.0.0",
     "mri": "^1.2.0",

packages/mcp/package.json

@@ -31,7 +31,7 @@
     "test:types": "tsc --noEmit --skipLibCheck"
   },
   "dependencies": {
-    "codefetch-sdk": "^2.0.0",
+    "codefetch-sdk": "workspace:*",
     "@modelcontextprotocol/sdk": "^1.13.2"
   }
 }

scripts/release-fix-workspace.ts

@@ -355,9 +355,7 @@ async function publishPackages() {
           `## ${mainVersion.replace(/\./g, "\\.")}([\\s\\S]*?)(?=## |$)`
         );
         const match = changelog.match(versionSectionRegex);
-        let releaseNotes = match
-          ? match[1].trim()
-          : `Release ${mainVersion}`;
+        let releaseNotes = match ? match[1].trim() : `Release ${mainVersion}`;
 
         // Clean up the release notes (remove leading/trailing whitespace)
         releaseNotes = releaseNotes.replace(/^\n+|\n+$/g, "");