This release helps PMs move faster in multi-step AI workflows by making follow-up chats more context-aware and easier to trust. You can quickly confirm what the model received for a turn, keep loaded skills and org knowledge reliably available across turns, and avoid duplicate or misplaced project links by checking where a URL is already referenced before creating new work.

• Improved LLM auditability in debug mode by combining preflight payload capture (including the fully composed systemPrompt and messagesForAgent) with step-level request/response logging, making Inspect Turn easier to trust when validating model behavior.
• Improved cross-turn memory reliability by carrying forward compact prior tool outcomes and preserving version-aware loaded context in LOADED_CONTEXT_IN_THIS_CHAT (including updated_at freshness checks), so follow-up turns retain loaded skills and org knowledge without replaying verbose tool JSON.
• Fixed context window usage calculation to avoid double counting across a chat session: context percentage and context-attachment safety checks now use current-turn context input usage (plus active captured context), not cumulative lifetime session tokens.
• Improved inline captured-context tokens in the chat composer to use readable page/screen labels (for example [[screen:2026 TikTok...]]) instead of opaque short IDs.
• Improved edit-event cards in chat and Edit History with clearer no-border visual separation and more readable add/remove change highlighting.
• Replaced the sidepanel header Home and Settings buttons with an options quick-nav dropdown that links directly to every options page (Chats, Skills, Prompts, Org Knowledge, Edit History, and Settings), now using DaisyUI's native dropdown behavior for more consistent open/close interactions and fewer custom-state edge cases.
• Removed automatic sidepanel focus jumps (on open, send completion, loading completion, and session/context updates) so the project picker no longer briefly opens and accidental project selection is less likely when you refocus the panel.
• Expanded the local search_projects agent tool with richer filtering and sorting (sort_by, sort_order, status, priority range, tags, recency windows like last 7 days, and stale-project queries), plus clearer parameter guidance so project triage requests resolve more reliably in chat.
• Added a Skills page search box with instant local filtering by skill name/description and a clear no-results state, plus improved input surface contrast so the field is visually distinct in light theme.
• Added strict duplicate-link protection to add_project_link: the tool now canonicalizes URLs and returns a clear error when the exact same link (including params) is already attached to a project.
• Added a new find_projects_by_link tool so the assistant can answer “is this URL already linked anywhere?” and return all matching projects plus the matched link records

This release gives PMs better control over both reasoning depth and captured page context. You can now tune how much thinking effort the assistant spends on harder work directly in chat, and add highlighted text to context explicitly with a keyboard shortcut instead of having context appear unexpectedly when the composer gains focus. This makes quote capture from customer calls, competitor pages, and docs more intentional and easier to trust.

• Added model-level reasoning controls in the sidepanel chat model popup so you can set reasoning depth inline (low, medium, high) while selecting your model.
• Added support flags in the model catalog (supports_thinking) so reasoning controls only appear when the currently selected model supports thinking, while still showing capability hints next to compatible models.
• Added provider-specific reasoning settings at runtime for OpenAI, Anthropic, and Google models so the selected reasoning depth is actually applied to API calls.
• Improved Google/Gemini reasoning behavior by requesting thought content explicitly (includeThoughts) so thinking output can be surfaced in the chat reasoning UI when the provider returns it.
• Added persistent reasoning-level preference in extension AI settings so your selected depth carries across sessions instead of resetting each time.
• Added a native highlighted-text capture shortcut (Ctrl+Shift+Y / Command+Shift+Y) that opens the sidepanel and adds the active selection to chat context in both Chrome and Firefox.
• Replaced legacy focus-based selected-text auto-attach behavior with explicit shortcut capture, so highlighted text is only added when the user deliberately triggers it.
• Added clearer in-panel capture feedback for shortcut flows, including no-selection and permission guidance, and deduped repeated captures of the same selection (url + exact text) in a chat.

0.0.10
- Fixed a bug where the AI agent's first edit_document or replace_document call would always succeed but any subsequent edit in the same conversation would fail with a spurious conflict error. The root cause was a timestamp mismatch: the document's updated_at and the edit event's timestamp were generated separately (at slightly different moments), so the concurrency check on the second call would compare two different timestamps and reject it. The tool now returns the document's actual updated_at so the agent always has the correct token for chained edits.
- Fixed a related bug where AI document edits would always conflict for new users (or after a data reset) because the default org knowledge document was never persisted to storage. Each read generated a fresh in-memory document with a new timestamp, so the timestamp in the system prompt never matched the timestamp seen during the edit. The default document is now written to storage on first read, giving it a stable updated_at.
- The open_web_pages agent tool now accepts an optional target parameter ("new_tab" or "current_tab"). When "current_tab" is chosen, the agent navigates the active browser tab to the URL instead of opening a new one — useful when the user says "go to X" rather than "open X in a new tab". Additional URLs beyond the first still open in new tabs.
- SupaPM now publishes a CycloneDX Software Bill of Materials for the browser extension on the main site, linked from the footer, so security and compliance teams can review the extension dependency supply chain before approval.
- Highlighted-text context now uses explicit optional site permissions. SupaPM asks before enabling selected-text capture for a site, supports either "this site" or "all sites", and lists granted page-context permissions in Settings so users can remove them.
- Removed the floating "Add as context" button from webpages. When page context is enabled for a site, highlighted text is added as a context chip when the user focuses the sidepanel chat input.

v0.0.9 has the below updates

• AI model catalog: The extension now discovers supported providers and models from SupaPM's public model catalog, with versioned catalog responses and automatic filtering for unavailable or incompatible models.
• New model options: Updated the catalog with Claude Opus 4.7, Claude Sonnet 4.6, Claude Haiku 4.5, Gemini 3.1 Pro Preview, Gemini 3 Flash Preview, and Gemini 3.1 Flash-Lite Preview.
• Manual model configuration: Settings now support a manual OpenAI-compatible endpoint for local or custom-hosted models, while keeping the SupaPM catalog as the recommended default.
• Chat model controls: Chat surfaces now show the active model, allow catalog model switching, and display token/context usage metadata when available.
• Chat tool context cache: Loaded Org Knowledge sections and Skills are now reused across turns, reducing repeat tool calls while surfacing cached-token diagnostics.
• Chat reasoning display: Responses with explicit <think> reasoning blocks now show that reasoning in a collapsible Thinking row, styled consistently with tool-call activity. The same row appears while chat is waiting for the first model response.

v0.0.8 Updates
- Skills don't use modals for editing any more - they have dedicated pages like projects - easier for editing
- Skills also now use the markdown editor instead of just raw markdown - a better experience.
- UI for chat is updated slightly to move the input into the actual column where the chat conversation lives
- Auto-scroll only follows new output when you’re already at the bottom.
- Removed snap points for smoother, more natural vertical scrolling.
- Notion-style @ date mentions (e.g. today, tomorrow, next week), keyboard menu support, date picker editing, and @date(YYYY-MM-DD) round-trip support.

v0.0.7 Updates
- Improved project selection and management - no longer auto-switching projects
- LLM now has new abilities to edit skills, org knowledge, and projects - WITH an audit log and revertability of changes
- Added edit history page so you can see, per document, your changes and the agent's changes
- Project links now show you (via opacity) when you're on that page
- Favicon fixes for project links (subdomains weren't working)

0.0.6
Markdown tables: Adding the ability to create tables in the Projects makes structuring things a little better.
--
Side panel / layout: Two-column mode when the main layout region is ≥ ~700px wide: project and chat sit side by side; below that width, a vertical stack keeps chat depth for long replies. Width uses a ResizeObserver (more reliable in Chrome’s side panel than container-query-only rules here). Horizontal overflow from w-full plus horizontal margins on the content wrapper is fixed; the shell adds min-w-0 / overflow-x-hidden so wide chat content is less likely to force sideways scroll.
--
Side panel / projects: Reworked project search and selection — the combobox lives in the header to save vertical space, stretches to use available width between the title and actions, and its results dropdown stacks correctly above the project card. Project details use DaisyUI’s checkbox collapse for expand/collapse, with scrolling and padding fixes so long project content stays usable. Still a bit rough around the edges, but noticeably better than before.
--
Projects / links: Favicons on project links (side panel and options) so destinations are easier to scan at a glance, with improved resolution/fetching when links are added or updated.
--
Chat: Assistant messages with fenced code blocks now show each block in a dedicated panel with a Copy button, so you can grab snippets without fighting partial selection in rendered markdown. Plain text and inline formatting in the same message remain easy to select as usual.
--
Theme: Fixed icons so they display correctly in both dark mode and light mode (contrast, assets, and theme-aware styling where needed).

Added skills! SupaPM can now be given clear instructions for common tasks and workflows like creating PRD's, generating status updates, and lots, lots more...