adrian

Hi @androbey - I haven't seen any issues like that, but I wonder if a modules > refresh and a hard reload in the browser might fix it? Also, what version did you upgrade from when it was working ok?

Maybe it would be helpful to have an AI prompt (like Ryan's Agent Tools) built-in to the Console panel so you can prompt your way to a script, or ask it to fix/extend an existing script?

Hi everyone, in particular @ryan @Peter Knight@ukyo @gebeer @maximus who seem to have been most AI active lately. I've just added dai() and bdai() dumping calls so that objects and arrays are rendered in plain text format more friendly to LLMs, but I am curious what AI/LLM integration features you think would be most useful? Claude suggested an MCP server - here is its plan. Does this sound useful? Any other ideas? Two processes, loosely coupled: ┌──────────────────┐ ┌──────────────────────────┐ │ Claude / Cursor │ stdio MCP │ TracyDebugger site │ │ client │ ─────────────────► │ │ │ │ HTTP + token │ tracy-ai/* endpoints │ └──────────────────┘ ◄───────────────── └──────────────────────────┘ MCP server — a tiny program the agent launches over stdio (the MCP transport). Ships as a sibling module (TracyDebuggerMCP/) or a standalone npm package the user npx's. TracyDebugger HTTP endpoints — new authenticated tracy-ai/* routes inside the ProcessWire site. The MCP server is just a thin translator between MCP tool calls and these HTTP requests. The MCP server holds no site logic. It's a dumb adapter. All the real work (reading panels, redacting secrets, rendering plaintext) stays inside TracyDebugger where the ProcessWire API is available. What the agent sees A handful of tools in the MCP catalog: tracy_export_bundle(preset: "debug" | "performance" | "template" | "full") tracy_get_request_info() tracy_get_last_errors(limit: int = 10) tracy_get_slow_queries(limit: int = 10) tracy_get_template_schema(template: string) tracy_list_dumps() tracy_run_console(code: string) ← gated, opt-in only Every tool returns the scrubbed plaintext/JSON produced by AIExport — same output Phase 2's "Copy" button produces. Config on the site New module-config section: aiExportHTTPEndpointEnabled (default off) aiExportMCPToken — a random token generated once per site, shown to the user to paste into their MCP client config aiExportAllowConsoleExec (default off) — gates tracy_run_console aiExportAllowedIPs — optional whitelist Config on the client User's ~/.config/claude/mcp.json or equivalent: json { "mcpServers": { "tracy": { "command": "npx", "args": ["-y", "tracy-mcp"], "env": { "TRACY_URL": "https://mysite.test", "TRACY_TOKEN": "<paste token from module config>" } } } } The agent launches the MCP server locally; the MCP server talks to the site over HTTPS with the token. Auth Per-site token (generated in module config, rotateable). Token sent as Authorization: Bearer … header on every HTTP call. Optional IP whitelist on the site side. tracy_run_console additionally requires aiExportAllowConsoleExec=true — otherwise the MCP server gets a 403 and reports "console execution disabled for this site" to the agent. Example flow — agent debugging an error User in Claude: "Why is /about/team throwing a 500?" Agent: Calls tracy_export_bundle(preset: "debug"). MCP server hits GET https://mysite.test/tracy-ai/export?bundle=debug with the bearer token. Site responds with scrubbed JSON: request info, PW info, last error with stack, slow queries, recent PW logs. Agent reads the traceback, sees TemplateFile.php:123 Undefined index "featured_image", asks tracy_get_template_schema(template: "team"). Site responds with the template's fields — no featured_image field exists. Agent suggests the fix, possibly calls tracy_run_console (if enabled) to verify. No human pasting. Agent pulls what it needs on demand, scoped by the tool it calls. What ships where In TracyDebugger itself: the tracy-ai/* HTTP endpoints + auth + token config + AIExport (already built in Phase 1, extended in Phase 2). In the MCP server (separate repo, ~200 lines): tool definitions, HTTP calls, response shaping for MCP. This separation matters because the MCP server can be installed independently of the site, and the site is still useful without it (you can hit tracy-ai/export with curl directly). Footprint on production Zero unless you explicitly enable it. The endpoints, token, and MCP config are all opt-in behind module settings. That's the shape. The main design choices worth confirming before building: Token-only auth, or also require the existing Tracy access? — i.e., should the agent's token have to belong to an allowed Tracy dev user, or is a separate machine token fine? I'd lean separate machine token for agents; reusing session auth is awkward over stdio. Read-only by default? — I strongly recommend yes. tracy_run_console is the only write path and should be a separate opt-in. Does the MCP server live in this repo or a separate repo? — I'd say separate. Different language, different release cadence, and the site works without it.

- You can now cancel a long running console panel script. - There are new dai() and bdai() methods which dump the contents of objects etc in a plain text format which is more friendly for consumption by LLMs.

A few tweaks and a major improvement to the d() and db() calls from the Console panel - these now make use of Tracy's Lazy loading option so the DOM is not populated with huge nested objects - it gets generated dynamically as you toggle open each element with the object. This should solve the massive browser slowdown I am sure we've all experienced at times when dumping lots of large objects. This together with the change from localStorage to IndexedDB has made a huge improvement to the Console panel.

Looking forwarding to trying out the memory feature @ryan Here's something of related interest: https://github.com/memvid/claude-brain I do love this description of the issue: You: "Remember that auth bug we fixed?" Claude: "I don't have memory of previous conversations." You: "We spent 3 hours on it yesterday" Claude: "I'd be happy to help debug from scratch!"

Hi @PWaddict - just wondering how things are working out for? Did you also try the new v2?

New v5 namespaced version is released:

@maximus - sorry, not really, but happy to accept a PR if you're willing.

Well, it's been about a year in the making, but v5 is finally available. I have upgraded a lot of sites to it now without issues, but I would still caution you to be ready to revert (or delete module) files if the namespace changes cause any issues - there were a lot early on. The two new banner features are: the console panel can now run very long running scripts (there is a one hour limit just so that broken scripts don't run forever) which is great for massive batch modifications or the like. The dumps recorder panel (either manually loaded or via Enable Guest Dumps) now live polls for new dumps so you don't need to continually load the page to see the entries as they are logged. Have fun! Breaking Changes - Minimum requirements bumped to ProcessWire 3 and PHP 7.1 - Removed legacy Tracy 2.5.x core branch and FireLogger support - Panel DOM IDs now include `ProcessWire-` prefix — update any custom CSS/JS targeting panel IDs Namespace Support - Full `namespace ProcessWire` support across all panels and POST processing files - Autoloader bridge for seamless non-namespaced to namespaced module migration - Third-party panels bridged automatically via `class_alias()` Security - Comprehensive security hardening: XSS sanitization, CSRF protection on all panels, directory traversal fixes, CSP nonces on all inline scripts, cookie SameSite enforcement, and input sanitization New Features - Console — long-running scripts automatically switch to background polling, surviving gateway timeouts; session locks released so you can continue browsing while scripts run; storage migrated to IndexedDB - PW Version Switcher — extracted into its own class with automatic version reverting on failure - Dumps Recorder live polling — live-polls for new dumps from other users and guest sessions - File Editor BlueScreen integration — exception page links open in the built-in file editor - Various smaller additions across Diagnostics, API Explorer, Request Info, and PW Info panels Bug Fixes - PHP 8.x compatibility fixes for `htmlspecialchars()`, `trim()`, and `isset` null handling - Fixed Console panel snippet and polling issues, including cache-busting for CDN/proxy environments - Fixed File Editor not opening files linked from BlueScreen exception pages - Fixed Adminer URL/namespace issues and thumbnail viewer path handling - Fixed Debug Mode panel to use modern API methods instead of deprecated ones - Fixed API Explorer "What's New" section and reflection errors for hooked methods - Fixed "unsaved changes" false positive when saving pages in PW admin - Windows path and line ending fixes Performance - Session lock contention reduced - Various loop and query optimizations

Thanks Ryan - just starting to play with this now. Remember that for some reason Claude needs to be reminded to add a version check around around curl_close Deprecated: Function curl_close() is deprecated since 8.5, as it has no effect since PHP 8.0 in site/modules/AgentTools/AgentToolsEngineer.php:592

In case it helps anyone, these are the settings I've been using for while which I believe make a huge difference to how it looks and its usability: And then in custom CSS section: :root { --pw-text-color: #444; --pw-muted-color: #999; --pw-button-radius: 10px; --pw-main-background: #fbfbfb; --pw-menu-item-background-hover: #EEEEEE; } html { font-size: 15px; } .PageList .PageListItem:hover { box-shadow: inset 0 0.5px 0 0 #ccc, inset 0 -0.5px 0 0 #ccc; } h1, .uk-h1 { font-size: 1.6rem; } .uk-card :is(a, .uk-link) { text-decoration: none; } .InputfieldHeaderHidden { --pw-border-color: transparent }

211 71 46 Keep in mind that most of these are the parent/child templates for page reference "tag" fields.

No plans on that from me - I was just seeing if Claude could do it - I think it would be something that would require core modifications and maybe modules that plugged into the core abstraction layer for each DB.

I actually played around and got Claude to build an DB abstraction layer for PW that would allow using Postgres etc. Seems like there is definitely potential on this front.

Sorry about that @PWaddict - both versions should be fixed now.

Hi @PWaddict - I have fixed that in 1.4.0 but I have also created a new 2.0 branch if you'd like to test it. It contains a lot of bug fixes along with removing the need for DOMDocument and eval() Please let me know if you find any issues in 2.0

If anyone is interested in being able to set script-src-attr to "none" on the frontend of their sites, the namespaced branch of Tracy now uses eventListeners everywhere - no more inline handlers.

Did you all notice that hovering menu items on their website preloads the content? Not sure that's a great idea. It makes things feel snappy, but it makes me think of Chrome's preload feature and its privacy concerns: https://www.malwarebytes.com/blog/product/2026/02/chrome-preloading-could-be-leaking-your-data-and-causing-problems-in-browser-guard

Just thought it's worth noting that on modern setups, that 64K limit shouldn't apply anymore. If you run: lsblk -f and the replace the "/dev/nvme0n1p1" with the appropriate partition in this: sudo tune2fs -l /dev/nvme0n1p1 | grep features It will return something like: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent 64bit flex_bg sparse_super large_file huge_file dir_nlink extra_isize metadata_csum If you see "dir_nlink" then the limit doesn't apply.

I am sure many of you have seen @Ex-user's comments about AI data centres in the in the ProcessWire 3.0.257 – Core updates thread. It seems like we have lost them from the community, but I do think a Pub topic about the darker side of AI is worthwhile. I have watched the video that was posted and here is another one worth a watch. I must admit that Claude code is making me much more productive, so it's a strange situation to be in. I do worry about all of the environmental and human health concerns posted in that other video and all of the societal ones in this video and so I think it's important that we are at least aware of these things when we talk about AI in ProcessWire. Let's keep the dialog respectful and productive.

Thanks @monollonom - my main concern is really the upgrade process from the non-namespaced to the new namespaced version. Until a couple of days ago, this would break the entire site, but hopefully that is being managed now.

Not really sure this is a tutorial, but if you're struggling with large PNG files, here is a nice little hook to compress them with pngquant. Obviously you need to install pngquant (https://pngquant.org/) first and exec() can't be disabled. $wire->addHookAfter('Pageimage::size', function(HookEvent $event) { $img = $event->return; // the resized Pageimage variation $path = $img->filename; // Only process PNGs if (strtolower(pathinfo($path, PATHINFO_EXTENSION)) !== 'png') return; // Skip if already optimized (optional marker file approach) $marker = $path . '.pngquant'; if (file_exists($marker)) return; // Run pngquant (overwrites in place, quality 65-80) $cmd = sprintf( 'pngquant --quality=65-80 --force --output %s -- %s 2>&1', escapeshellarg($path), escapeshellarg($path) ); exec($cmd, $output, $returnCode); // Mark as optimized so we don't re-process on subsequent calls if ($returnCode === 0 || $returnCode === 99) { // 99 = quality target not met, kept original touch($marker); } });

One of the reasons I no longer use ChatGPT for anything - I don't want to get political here, but IYKYK. I would love to boycott Google and Amazon completely as well. I do my best on these fronts, but it's basically impossible.