Fabric: baixar versions anteriors

Changes
PR #2244 by taljeon: fix(core): preserve the selected vendor model spelling
- Preserve the selected vendor's canonical model spelling for explicit vendor and vendor/model-prefix requests.
- Add regression tests for case-insensitively equal IDs with different vendor spellings and catalog order.

Changes
PR #2266 by jayervilha: fix: remove zero-width spaces and BOMs from write_nuclei_template_rule pattern
- Removed 53 zero-width spaces (U+200B) and 5 byte order marks (U+FEFF) from the write_nuclei_template_rule pattern; no other content changed.
- Restored 10 code fences that a leading zero-width space had stopped CommonMark from recognising.
PR #2267 by T145: feat: add WebWright as an OpenAI-compatible provider
- Added WebWright as an OpenAI-compatible provider.
- Added a changelog entry documenting the new WebWright provider.

Changes
PR #2265 by Owaid0723666: feat: add onomeo as an OpenAI-compatible provider
- Added onomeo as an OpenAI-compatible provider, registered in the ProviderMap with the base URL https://onomeo.com/v1.
- Added a matching TestCreateClient test case.

Changes
PR #2263 by aldoyh and ksylvan: Add Arabic localization file with application options and commands
- Added an Arabic localization file covering application options and commands.
- Expanded Bahrain Arabic translations across the CLI, setup flow, and error messages.
- Added Saudi Arabic translations that match the Bahrain locale coverage.
- Removed generic Arabic translations in favor of a regional fallback strategy, with tests for locale candidate resolution and the default Bahrain fallback order.
- Refactored the code structure for improved readability and maintainability.

Changes
PR #2261 by ksylvan: fix: harden REST server, template engine and plugins
- Reworked the template engine to resolve each token exactly once in a single linear scan, preventing variable values from being re-expanded into plugin or extension calls and eliminating quadratic processing time on long brace runs.
- Added loopback security middleware and JSON content-type enforcement to the REST and Ollama-compatible servers, so that, when no API key is set, these servers accept only loopback or configured CORS hosts and origins.
- Confined the file plugin to a root directory with os.Root, rejecting absolute, ~, .., and escaping symlink paths, and enforcing read size limits on special files such as /dev/zero.
- Restricted the fetch plugin and the Ollama image loader to public network addresses through a shared HTTP client with DNS-level address checks, a 30-second timeout, a 5-redirect limit, and no API key sent to image hosts.
- Added per-session locking with atomic writes, request body size limits (50 MiB), merged .env updates, and DOMPurify sanitization plus a Content-Security-Policy in the web UI to prevent data loss, memory exhaustion, and unsafe HTML in rendered output.
- Breaking: With no API key, the server gives 403 to a request whose Host is not loopback. To use a LAN address or another host name, add its origin with --cors-origins.
- Breaking: JSON endpoints need Content-Type: application/json. Other requests get 415. This includes the Ollama-compatible /api/chat: add -H 'Content-Type: application/json' to a curl -d call.
- Breaking: The file plugin reads only relative paths.
- Breaking: The template engine does not expand a variable value again. With x = "{{name}}", {{x}} gives {{name}}.
- Breaking: The fetch plugin and the Ollama image loader ignore HTTP_PROXY, and they refuse loopback, private and other non-public addresses. This also applies to the CLI. An Ollama image URL on a LAN host now gives an error.
- Breaking: Fabric removes terminal control sequences (other than colors) from the CLI output. GET /config shows a fixed mask in place of each API key. After the upgrade, get the configuration again before you send it to POST /config/update. The server keeps an old mask, such as ****abcd, as a new key value.
- Breaking: The Streamlit UI binds to 127.0.0.1. To open it from a different computer, start it with --server.address 0.0.0.0 and set FABRIC_UI_TOKEN.
- Breaking: In a workflow, --input-has-vars applies only to the first input and to each input from the workflow file. The output of a step goes to the next step as literal text.
- code2context and the pattern update (-U) skip symbolic links.
- New session and context files get mode 0600, and new storage folders get mode 0700. Existing files and folders keep their mode.
- --notification-command gets the title and the message with no ", no ' and no \. POST /config/update merges into .env and answers 400 for CR, LF or NUL. A strategy name must obey the storage name rules, which refuse Windows device names such as con and aux on all platforms.
- Breaking: The REST server runs only the text and datetime template plugins in a pattern. A sys, file or fetch plugin token or an extension token in a pattern gives an error on /patterns/:name/apply and /chat. The CLI runs all plugins.
- Breaking: With no API key, the server accepts a POST, PUT, DELETE or PATCH request with an Origin header only from the same origin (host and port), a --cors-origins entry or a JSON request from the web UI on port 5173 or 4173 (vite preview). Before, it accepted each loopback origin on any port.

Changes
PR #2260 by Felixkw12: feat: add Opper as an OpenAI-compatible provider
- Added Opper as an OpenAI-compatible provider, registered in the ProviderMap with the base URL https://api.opper.ai/v3/compat.
- Listed Opper in the OpenAI-Compatible Providers section of README.md and README.zh.md.
- Added a matching TestCreateClient test case.

Changes
PR #2258 by JuampiHernandez: feat: add Firecrawl web search via --firecrawl_search flag
- Feat: add Firecrawl web search via --firecrawl_search flag

Changes
PR #2257 by ksylvan: feat: add DemonRoute as an OpenAI-compatible AI provider
- Added DemonRoute as an OpenAI-compatible AI provider.
- Registered DemonRoute with its OpenAI-compatible API base URL.
- Added DemonRoute to the README’s list of supported AI providers.

Changes
PR #2107 by johnpippett: Security: fix remote file write and shell injection in template extensions
- Fixed a remote arbitrary file write vulnerability by disabling automatic application of file changes when running in API mode, preventing unsupervised writes triggered through the REST API's PatternName input.
- Replaced the weak strings.Contains(path, "..") check in the file manager with proper path containment validation that rejects absolute paths and confirms resolved paths stay within the project root.
- Hardened the path containment check in ApplyFileChanges by using filepath.Rel, correctly handling edge cases such as a project root of /.
- Consolidated path validation in ParseFileChanges and ApplyFileChanges to a single filepath.IsLocal call, which rejects absolute paths, empty paths, directory traversal, and Windows reserved names.
- Added a regression test that fails if the path containment guard is removed.
PR #2187 by moritzschmitz-oviva: feat: add Eisenhower Matrix prioritization pattern
- Added a new Eisenhower Matrix prioritization pattern for task triage and decision-making.
- Renamed the pattern from eisenhower to eisenhower_matrix and removed the old pattern directory.
- Relocated the Eisenhower Matrix pattern into data/patterns.
- Registered the pattern's description and metadata with the BUSINESS, SELF, and STRATEGY tags.
- Added task classification and a weekly planning extract to the pattern.

Changes
PR #2201 by ghrom: feat: refresh Ultimate Law patterns to current doctrine, add judge_ultimate_law
- Add judge_ultimate_law pattern: derives ethical verdicts by applying the Ultimate Law framework's executable rulebook stratum by stratum — deny-by-default consent, derivation chains, integrity constraints, and a falsifiability line naming the fact that would flip each verdict.
- Refresh ultimate_law_safety to the framework's current doctrine: the Law quoted verbatim, consent channels, and new Forfeiture and Theft-by-withholding definitions (proportionality measured in kinds of harm; inability is never theft).
- Refresh audit_consent with the consent-channel threshold test (words, conduct carrying intention, agreement, or prior permission — otherwise there is no consent to audit) and doctrine-grounded revocability.
- Register the new pattern in pattern descriptions and extracts JSON files, categorize it under ANALYSIS and CR THINKING in suggest_pattern, and update pattern_explanations.md with renumbered entries.
PR #2223 by googio and ksylvan: feat: add Serply web search via --serply_search flag
- Added an optional Serply web search plugin (SERPLY_API_KEY) exposed through the new --serply_search flag, which runs a Google search and appends the results to the chat as markdown alongside the existing Jina scrape_question tool. Installations without configuration remain unchanged.
- Refactored the Serply integration to use a shared HTTP client with a 30-second timeout, removing per-instance client configuration.
- Simplified Serply request setup and removed the custom Fabric User-Agent header so requests use the default HTTP User-Agent.
- Documented the --serply_search flag in the README and added it to the shell completions.
- Updated search tests to use the shared HTTP client and dropped the obsolete User-Agent assertion.
PR #2253 by ksylvan: feat: expand pattern suggestions with Chinese workflows
- Added five Chinese-language workflow patterns across the relevant suggestion categories, covering summaries, reviews, translation, and poetry.
- Introduced deconstructive, rhetorical, and philosophical analytical lenses for text analysis.
- Documented eight additional patterns in the suggestion catalog, including descriptions, category tags, and prompt extracts.
- Renumbered subsequent catalog entries to accommodate the eight new additions.
- Updated the changelog with the entry for this release.

Changes
PR #2252 by ksylvan: feat: add configurable CORS support to REST and Ollama servers
- Added configurable CORS support to the REST and Ollama servers, with origins set through repeatable flags and environment variables.
- Applied shared CORS middleware across both the REST and Ollama servers.
- Required API keys for wildcard origins and excluded null origins for tighter security.
- Handled preflight requests before authentication and removed hardcoded chat origins.
- Loaded server settings from .env while preserving explicit overrides.
Direct commits
- Merge branch 'main' into pr/pattern-input-normalization
- Merge branch 'main' into fix/readme-yt-helper

Changes
PR #2173 by OdinKral and ksylvan: fix: clearer error when binary name used as pattern fallback
- Added a clearer error hint when a pattern is derived from the binary name, explaining that the executable name was used as a fallback and how to specify a pattern explicitly with -p.
- Introduced a patternFromBinaryName() helper and explicit tracking of binary-derived pattern names during flag initialization, so fallback hints appear only when a pattern truly came from the binary name.
- Prevented fabric-ai from automatically selecting a pattern by binary name by centralizing the list of executable names that do not trigger pattern selection.
- Added the pattern_from_binary_name_hint key with localized text across all ten non-English locale files for a consistent i18n bundle.
- Simplified error wrapping, removed obsolete detection tests, and added coverage for default executable exclusion and extension-stripped pattern selection.
Direct commits
- Merge branch 'main' into docs/sync-chinese-readme

Changes
PR #2225 by aiapienthusiast: feat(providers): add Cheaper Inference as an OpenAI-compatible provider
- Feat(providers): add Cheaper Inference as an OpenAI-compatible provider

Changes
PR #2239 by ksylvan: feat: add llama.cpp support through the LM Studio-compatible client
- Added llama.cpp support through the LM Studio-compatible client, registered with a localhost endpoint and an optional API key.
- Replaced dots with underscores in environment variable names so dotted plugin names resolve correctly.
- Added tests covering dotted plugin names to confirm the expected environment prefixes.

Changes
PR #2204 by ericcurtin: feat: add llmman to the OpenAI-compatible providers
- Added llmman as an OpenAI-compatible provider, enabling local models distributed as OCI artifacts to be served through an OpenAI-compatible API on port 17434.
- Implemented the integration with a single ProviderMap entry, following the LiteLLM precedent for a localhost BaseURL and requiring no special handling.
- Extended the existing TestCreateClient table with a new case covering the llmman provider.

Changes
PR #2212 by pacocartones: fix(ollama): honor caller context in SendStream and ListModels
- Fixed SendStream and ListModels in the Ollama vendor to honor the caller's context.Context instead of substituting context.Background(), so cancellations now propagate to o.client.Chat and o.client.List.
- Resolved an issue where disconnecting web clients left Ollama generations running in the SSE server path, wasting compute and leaking in-flight requests (closes #2196).
- Added a table-driven httptest regression test that cancels mid-stream and asserts SendStream returns context.Canceled promptly without draining the full stream.
- Simplified the Ollama stream cancellation test coverage, including streamlined streaming responses and server cancellation checks.
- Replaced manual cancellation error assertions with require.ErrorIs checks and removed explicit timeout guards and stream completion tracking.

Changes
PR #2236 by ksylvan: chore: clarify comments and remove obsolete code examples across backend and web
- Removed redundant comments across Go sources and Svelte components.
- Clarified documentation for provider authentication, streaming, and model discovery.
- Documented cache behavior, configuration precedence, and locale fallback rules.
- Added notes on security safeguards, implementation limitations, and existing test boundaries.
- Deleted obsolete commented-out code without altering runtime behavior.

Changes
PR #2235 by ksylvan: fix: honor configured yt-dlp arguments for REST API transcripts
- Fixed the REST API so that configured yt-dlp arguments are honored for transcript requests.
- Passed flag and configuration arguments through to the YouTube plugin.
- Applied plugin arguments to both plain and timestamped transcripts.
- Added server wiring tests to verify yt-dlp argument propagation.
- Documented configuration support for the REST transcript endpoint.

Changes
PR #2234 by ksylvan: chore: update web dependencies and refresh npm and pnpm lockfiles
- Upgraded Svelte to 5.57.1 and SvelteKit to 2.70.3.
- Bumped Vite to 8.3.1 and Vitest to 4.1.11.
- Updated Skeleton to 5.0.1 and Zag to 1.43.0.
- Upgraded TypeScript ESLint, Svelte linting, and formatting tools.
- Refreshed both lockfiles, updated Rolldown bindings, and removed WebAssembly runtime dependencies.

Changes
PR #2219 by anandghegde: feat: add --extract and --extract-last to output only a fenced code block
- Feat: add --extract and --extract-last flags that output only the first or last fenced code block of the response, like llm -x
- Falls back to the full response when no fenced code block is found
Direct commits
- Merge branch 'main' into docs/sync-chinese-translation

Changes
PR #2147 by medhwu: feat: add Chinese AI vendors, i18n translation fixes, and 5 Chinese patterns
- Added OpenAI-compatible provider support for three Chinese AI vendors: Aliyun DashScope (Tongyi Qianwen), Zhipu AI (bigmodel.cn/GLM), and ByteDance Ark (Doubao).
- Added test coverage for all three new Chinese AI vendor providers.
- Improved Chinese (i18n) translation quality by unifying "provider" terminology across 24 entries and correcting residual English text.
- Introduced five new Chinese-language patterns: article summarization, contract risk review, classical poetry analysis, English-to-Chinese news translation, and code review.

Changes
PR #2069 by 1rashiid and ksylvan: feat(cli): add --workflow for sequential pattern composition
- Added a --workflow flag that chains multiple patterns together from a YAML or JSON file, piping each step's output into the next step's input.
- Preserved compatibility with existing flags, including --stream, --dry-run, -m/-V, --context, --strategy, and --language, while leaving the single-pattern -p path untouched.
- Added pre-flight validation that catches empty steps, missing or repeated patterns, and unknown pattern names before any LLM call is made.
- Added per-step overrides for input, model, vendor, and variables, with progress and error messages prefixed as [step N/TOTAL pattern] on stderr to keep stdout pipe-clean.
- Added shell completions for --workflow in Bash, Zsh, and Fish, including workflow file path suggestions.

Changes
PR #2229 by ksylvan: Show provider error details in Codex 401 responses
- Preserved Codex provider details in authentication errors, appending them after the localized login error message.
- Retained the underlying authentication errors so they can be inspected through unwrapping.
- Added regression test coverage for Codex 401 provider messages.
- Removed the unused calculateFileHash method from the extension registry, along with its crypto/sha256, encoding/hex, and io imports.
- Removed the unused Verify and ListExtensions methods from the extension registry.

Changes
PR #2228 by ksylvan: refactor: simplify Ollama chat prompt building and error replies
- Refactored Ollama prompt construction to unify message joining using strings.Builder.
- Centralized error handling and responses while preserving streaming and JSON output formats.
- Added tests covering streaming errors and multiple-message forwarding.

Changes
PR #2226 by ksylvan: feat: add Claude Code provider through the local CLI
- Added a Claude Code CLI provider that authenticates through an existing local subscription login.
- Added support for streaming responses, model selection, and configurable thinking effort.
- Added handling for local images and base64 attachments via temporary files.
- Stripped Anthropic environment variables to preserve subscription-based billing.
- Added Claude Opus 5.5 with extended context support.

Changes
PR #2224 by ksylvan: chore: remove GitHub Models provider integration and documentation
- Removed the GitHub Models provider from the compatible provider registry, dropping the integration entirely.
- Dropped GitHub API version headers from model discovery requests.
- Replaced GitHub cache fixtures with generic test providers and removed GitHub header verification along with its supporting test utilities.
- Deleted the GitHub Models setup documentation and the Copilot cross-reference.
- Removed GitHub Models listings from both the English and Chinese READMEs.

Changes
PR #2220 by jiweiyeah: feat(providers): add Y-API as an OpenAI-compatible provider
- Feat(providers): add Y-API as an OpenAI-compatible provider

Changes
PR #2216 by ctbaum: fix: enable raw mode for GPT-6 models
- Fix: enable raw mode for GPT-6 models

Changes
PR #2211 by ksylvan: fix: prevent pattern loader temporary directory leaks
- Prevent pattern loader leaks by lazily creating temporary directories during database population and cleaning them up after successful or failed downloads.
- Add regression tests for lazy directory creation and cleanup.

Changes
PR #2209 by kadiryildiz283: feat(i18n): add Turkish (tr) translation
- Feat(i18n): add Turkish (tr) translation

Changes
PR #2206 by ksylvan: fix: confine storage names and authenticate Ollama serve
- Reject unsafe cross-platform storage names and directory traversal attempts.
- Confine symlink targets to configured filesystem storage directories.
- Require API keys for non-loopback server bindings and authenticate Ollama routes.
- Validate chat pattern, context, and session names early while preventing internal filesystem details from leaking through client errors.
- Default the REST server to loopback port 8080 and add regression coverage for traversal, symlink, and authentication security.

Changes
PR #2203 by pacocartones and ksylvan: fix(web): preserve multi-byte UTF-8 split across SSE chunks in chat stream
- Fix: preserve split UTF-8 characters in streaming responses
- Decode API response chunks with persistent streaming state
- Reuse streaming decoder when inspecting chat backend events

Changes
GitHub unexpected status code: 410 from provider GitHub, response body: {"error":{"code":"github_models_retirement_brownout","message":"GitHub Models is temporarily unavailable as part of a scheduled retirement brownout."}}
PR #2200 by ksylvan: fix: persist Codex OAuth tokens after refresh
- Persist refreshed Codex OAuth tokens securely across processes, writing rotated credentials atomically with cross-process locks.
- Reload stored tokens before refresh and return valid in-memory tokens first, preventing stale disk values from overwriting fresh credentials.
- Reuse valid refresh tokens before launching interactive authentication, requiring account identifiers and unexpired tokens for fast-path reuse.
- Preserve existing environment values, save environment files atomically while holding locks, and enforce secret-only permissions.
- Reactivate configured vendors and surface configuration and provider failures immediately, with localized Codex prompts and persistence errors.

Changes
PR #2186 by giodamelio: Update Nixpkgs version for newer Go version
- Updated the Nixpkgs version to provide a Go release that satisfies the go.mod requirement of Go >= 1.26.0, as the previously pinned Nixpkgs only shipped Go 1.25.4.

Changes
PR #2185 by ksylvan: feat(web): replace PDF.js pipeline with pdf-inspector WASM worker
- Replaced the PDF.js processing pipeline with the new pdf-inspector.worker.ts, backed by @firecrawl/pdf-inspector-wasm. The worker initializes WASM one time and processes transferred ArrayBuffers.
- Added worker error handling. A worker crash now rejects all pending requests with a clear error and terminates the worker. The next request starts a new worker.
- Fixed the file attachment behavior. The app now parses each file and stores its content without a chat request. One submit sends exactly one streamChat call, not one call for each attached file.
- Removed the legacy PDF dependency chain: pdf-to-markdown-core, pdfjs-dist, pdf-config.ts, and the related build configuration. @firecrawl/pdf-inspector-wasm is now the only PDF dependency.
- Simplified the worker boundary in follow-up refactors: merged redundant methods, removed dead code, and inlined a single-use interface. The full test suite continued to pass.

Changes
PR #2182 by drawliin: fix(ollama): close stream channel on errors
- Fix(ollama): close stream channel on errors

Changes
PR #2171 by OdinKral: feat(scripts): add pattern/maintenance audit script
- Adds scripts/audit-patterns.sh, a dependency-free audit tool for the patterns library that detects thin patterns (under 15 lines), bloated patterns (over 50 KB), stale hardcoded model references, missing INPUT sections, i18n key gaps, and shell completion gaps. The script is read-only and exits with code 0 by default, or 1 when run with --strict for CI integration.

Changes
PR #2116 by ksylvan and dependabot: fix(web): repair the chat page and modernize the build toolchain
- Modernized the web stack by upgrading Tailwind, Skeleton, SvelteKit, Vite, and supporting dependencies.
- Replaced removed Skeleton components with compatible local implementations to restore UI functionality.
- Improved pre-stream chat error reporting to prevent duplication of streamed errors, and hardened vendor list parsing to return empty model arrays on malformed input.
- Fixed stream completion crashes, normalized displayed error messages, corrected toast stacking order, and added warning notification support.
- Restored linting, formatting, testing, and pnpm override configuration, and preserved custom themes through the Tailwind CSS-based configuration migration.
PR #2178 by OdinKral and ksylvan: feat(patterns): add generate_frontmatter for PKM/Obsidian users
- Add generate_frontmatter to conversion, extraction, and writing categories
- Document PKM-ready YAML metadata fields in pattern explanations
- Register pattern descriptions and extracts for suggestion workflows
- Credit both contributors in incoming changelog entry

Changes
PR #2180 by ksylvan: fix: include Grok in localized --search help text
- Fix: Updated all 11 locale entries in the i18n message catalog to include "Grok" in the enable_web_search_tool help text, ensuring Grok (xAI) web search support is discoverable via fabric --help. Also updates the generated README help block to match. No functional change.

Changes
PR #2168 by ksylvan: fix: complete localized setup and error messages for supported locales
- Fix: complete localized setup and error messages across supported locales
- Translate Bedrock setup prompts across nine supported locales
- Localize datetime and system template errors consistently
- Translate Persian Spotify errors and setup guidance
- Correct Japanese and Polish file operation log labels

Changes
PR #2123 by ksylvan and OdinKral: fix: block path traversal in pattern name lookup
- Security Fix: Blocked path traversal attacks in pattern name lookup (closes #2094) — pattern names containing .. could previously escape the patterns directory and read arbitrary files via filepath.Join; a guard has been added at the top of getFromDB, an i18n key pattern_invalid_name has been added to all 11 locale files, and test cases now cover all common traversal variants.
- New translations for the "invalid pattern" user-facing string.

Changes
PR #2166 by ksylvan: feat: add Claude Opus 5 support and refresh dependencies
- Add Claude Opus 5 to the supported model selection.
- Disable sampling parameters for Claude Opus 5 requests.
- Restrict one-million-token beta headers to compatible Claude models only.
- Remove unsupported 200K-context models from the beta header mapping.
- Upgrade Anthropic, AWS, Ollama, Google, and supporting dependencies.

Changes
PR #2135 by octo-patch: feat: upgrade MiniMax default model to M3
- Upgraded the MiniMax default model to M3, making it the new flagship selection.
- Added MiniMax-M3 to the static model list as the first entry, establishing it as the default.
- Retained MiniMax-M2.7 and MiniMax-M2.7-highspeed as available alternative models.
- Removed deprecated models (M2.5, M2.5-highspeed, M2.5-lightning, M2, M2.1, and M2.1-lightning) from the static list.

Changes
PR #2161 by ksylvan: fix: respect Anthropic chat option max token overrides
- Fix: respect Anthropic chat option max token overrides
- Use configured Anthropic max tokens as default
- Apply chat option max tokens when provided
- Preserve existing behavior for missing token overrides
- Add tests for default max token selection
- Add tests for explicit max token overrides
Direct commits
- Chore: clean up ChangeLog

Changes
PR #2155 by ksylvan: Claude Sonnet 5 Anthropic support
- Add Claude Sonnet 5 to supported Anthropic models
- Enable one-million-token context beta for Claude 5 models
- Omit sampling parameters for Claude Sonnet 5 requests
- Centralize Anthropic sampling restrictions behind prefix matching
- Remove older Claude 4 aliases from model listings
PR #2156 by ksylvan: Make it possible to back-fill missing ChangeLog entries
- Add support for changelog generation for closed pull requests via a new --closed-ok flag, bypassing open-state validation.
- Skip mergeability checks when processing closed pull requests to allow smooth back-filling of missing entries.
- Store the closed pull request allowance setting in the generator configuration for consistent behavior.
- Improve error messaging to guide users toward using --closed-ok when validation errors occur on closed PRs.
- Introduce the --closed-ok flag as the primary mechanism for enabling back-fill workflows on previously closed pull requests.

Changes
PR #2138 by ksylvan: New Claude Fable model + cache OpenAI model discovery and handle provider rate limits
- Add persistent cache for provider model discovery results, improving performance and reliability of provider integrations.
- Serve stale model caches during discovery failures, ensuring continued operation when upstream providers are unavailable.
- Add Claude Fable 5 Anthropic model support, with sampling parameters automatically omitted for compatibility.
- Return concise localized errors for rate-limited model fetches, with updated translations across all supported locales.
- Update Go dependencies for AI provider integrations to keep upstream libraries current.

Changes
PR #2136 by ksylvan: chore: extend sampling param exclusion to Opus 4.8 models
- Extends the sampling parameter exclusion logic to cover Opus 4.8 models, ensuring consistent behavior alongside the existing Opus 4.7 exclusion.
- Adds the claude-opus-4-8 model prefix to the sampling parameter exclusion check.
- Updates the associated code comment to explicitly reference Opus 4.8 models.

Changes
PR #2132 by ksylvan: Add Claude Opus 4.8 model and bump Go toolchain and dependencies
- Add Claude Opus 4.8 to the list of supported models.
- Upgrade the Go toolchain to version 1.26.0.
- Bump anthropic-sdk-go to v1.46.0.
- Update AWS SDK and Bedrock service modules to their latest versions.
- Bump the Ollama client to v0.24.0.

Changes
PR #2111 by ksylvan: fix: omit Anthropic sampling params for Claude Opus 4.7
- Fix: Omit Anthropic sampling parameters for Claude Opus 4.7 to ensure compatibility.
- Add a sampling parameter guard specifically for Opus 4.7.
- Omit temperature and top_p for models that do not support these parameters.
- Preserve existing TopP and temperature selection behavior for compatible models.
- Add unit test coverage for the Opus 4.7 sampling parameter omission.

Changes
PR #2079 by teamsincetoday: Add 3 commerce intelligence patterns: affiliate extraction, video entities, monetization
- Added extract_affiliate_products pattern to surface both sponsored and organic affiliate opportunities from any video transcript, going beyond the existing extract_sponsors pattern.
- Added extract_video_commerce_entities pattern to identify all commercial entities in video content, categorized by type, timestamp position, and purchase likelihood.
- Added analyze_monetization_opportunities pattern to map audience intent to revenue strategies, covering affiliate links, sponsorships, and digital products.
- Registered all three new patterns in pattern_descriptions.json and pattern_extracts.json with appropriate metadata and tags.
- Integrated all three patterns into suggest_pattern under the ANALYSIS, BUSINESS, and EXTRACT categories, and documented them in pattern_explanations.md.
PR #2086 by majiayu000: fix: parse vendor prefix from model name for vendor/model convention
- Fix: Added fallback logic to parse the vendor prefix from a model name when no vendor is explicitly specified. When a model string such as ollama/llama3 is passed, the lookup no longer fails with a "could not find vendor" error; instead, the first path segment is split and checked against known vendors, correctly resolving the model to llama3 under the Ollama vendor group.

Changes
Direct commits
- Fix: fall back to streamed delta text when completed Codex response is empty
- Prefer extracted completed text only when content stays non-empty
- Fall back to accumulated streamed delta text otherwise
- Preserve streamed response text before completed response evaluation
- Add regression test for empty completed output text
- Simulate SSE delta stream followed by blank completion
- Verify Send returns delta text when completion lacks content

Changes
PR #2093 by alecjmckanna: feat: add --readpattern flag to print pattern contents to terminal
- Adds a new --readpattern <name> CLI flag that prints the raw contents of a named pattern's system.md file to stdout, making it easy to inspect a pattern's instructions without navigating the filesystem manually.
- Custom pattern directories are respected: the user's custom patterns directory is checked first before falling back to the main patterns directory, consistent with all other pattern lookups.
Direct commits
- Docs: update Docker config mount path for appuser
- Replace container config mount path from root to appuser
- Align setup example with non-root container home directory
- Align pattern usage example with appuser config location
- Align REST API example with updated config mount target
- Update English and Chinese README Docker instructions consistently

Changes
PR #2091 by jimscard and ksylvan: Update Dockerfile for best practices and critical CVE fixes
- Pins Alpine 3.21 and Go 1.25.9 explicitly for reproducible, auditable builds.
- Installs the Go toolchain directly in the builder stage, removing the dependency on an unavailable upstream golang tag.
- Upgrades setuptools to remediate the critical vulnerability CVE-2025-47273.
- Refreshes the yt-dlp installation path to align with current packaging conventions.
- Configures the final image to run as a non-root user, reducing the container's attack surface.

Changes
PR #2075 by ksylvan and mikaelpr: refactor: extract OAuth and auth logic from Codex client module
- Refactored the Codex client module by extracting all OAuth and authentication logic into a dedicated module, improving separation of concerns.
- Removed the OAuth flow, PKCE handling, and token refresh logic from codex.go, streamlining the client's core responsibilities.
- Removed the auth transport round-trip retry logic, simplifying the HTTP transport layer.
- Removed JWT parsing and token expiry utilities, along with unused OAuth types and helper structs, reducing dead code in the package.
- Added a test for SendStream HTTP error mapping and channel close behavior, improving test coverage for the client module.

Changes
PR #2068 by dependabot and ksylvan: chore(deps): bump google.golang.org/grpc from 1.79.0 to 1.79.3 in the go_modules group across 1 directory
- Bumped google.golang.org/grpc from v1.79.0 to v1.79.3 as an indirect dependency update.
- Bumped anthropic-sdk-go from v1.23.0 to v1.27.1, keeping the Anthropic client library up to date.
- Removed deprecated ModelClaude4Sonnet20250514 and ModelClaude4Opus20250514 model aliases from the Go module.
- Updated gin-gonic/gin to v1.12.0 and go-git to v5.17.0, bringing in the latest framework improvements.
- Bumped ollama/ollama from v0.16.2 to v0.18.2, incorporating the latest Ollama client updates.

Changes
PR #2064 by JasonYeYuhe and ksylvan: docs: add Chinese translation (README.zh.md)
- Added a new Chinese translation of the README file (README.zh.md).
- Refined Chinese translations for improved idiomaticity and natural language flow.
- Removed a duplicate key in the Chinese (zh) locale file.

Changes
PR #2066 by octo-patch: feat: upgrade MiniMax default model to M2.7
- Upgraded the MiniMax default model to M2.7, the latest flagship model with enhanced reasoning and coding capabilities.
- Added MiniMax-M2.7 and MiniMax-M2.7-highspeed to the static model list.
- Placed M2.7 models at the top of the list as the new defaults.
- Retained all previous models (M2.5, M2.5-highspeed, M2.5-lightning, M2, M2.1, M2.1-lightning) as available alternatives.

Changes
PR #2056 by mikaelpr and ksylvan: feat: add Codex vendor with OpenAI OAuth
- Add new Codex AI vendor plugin with OpenAI OAuth PKCE flow
- Implement browser-based login with automatic token refresh on 401
- Allow explicit Codex model selection bypassing model listing
- Expose shared OpenAI BuildResponseParams and ExtractText helpers, and move system/developer messages into Codex instructions field
- Add comprehensive unit tests for Codex OAuth, streaming, and retry
- New internationalization keys added for all 11 languages.

Changes
PR #2060 by octo-patch: feat: add MiniMax-M2.5-highspeed to static model list
- Added the MiniMax-M2.5-highspeed model variant to the static MiniMax model list. This variant delivers the same performance as MiniMax-M2.5 but with faster inference speed.

Changes
PR #2058 by dependabot: chore(deps-dev): bump devalue from 5.6.3 to 5.6.4 in /web in the npm_and_yarn group across 1 directory
- Bumps the indirect dependency devalue from version 5.6.3 to 5.6.4 in the /web directory as part of the npm_and_yarn dependency group.

Changes
PR #2052 by praxstack and ksylvan: feat(bedrock): dynamic region fetching and AWS_PROFILE conflict fix
- Added dynamic Bedrock region fetching from botocore's endpoints.json, expanding support to 40+ regions instead of a hardcoded list of 6, with a static fallback on error.
- Fixed AWS_PROFILE environment variable conflict that caused 'failed to get shared config profile' errors for users who had AWS_PROFILE set for other tools (e.g., Terraform, AWS CLI) while using explicit access keys or static credentials.
- Improved empty auth choice handling to gracefully skip the selection instead of throwing an error.

Changes
PR #2054 by pretyflaco: fix(ollama): Map thinking levels to Ollama API and convert single system-role messages
- Fix(ollama): map thinking levels and convert single system-role messages
Direct commits
- Fix: Manual edit of changelog.db and fixes for ChangeLog

Changes
PR #2050 by ksylvan: Update web dependencies - fix npm override issue
- Chore: update web dependencies - fix npm override issue

Changes
PR #2044 by PrakharMNNIT and ksylvan: feat: Add Bedrock bearer token (ABSK) authentication and guided setup
- Adds 3-tier authentication for AWS Bedrock: bearer token (ABSK), static AWS credentials (access key + secret key), and the default AWS credential chain.
- Introduces a custom guided setup flow covering auth method selection, region, and model configuration.
- Removes the hasAWSCredentials() gate so that Bedrock always appears in the setup wizard.
- Applies a temperature-only fix (no top_p) for Claude models running on Bedrock, along with API key masking during re-setup.
- Adds a fallback model list for when the ListFoundationModels API is inaccessible, i18n support for 11 locales (3 new keys each), and 25 unit tests.
PR #2047 by bvandevliet: Explicitly instruct YT summary task to be aware of possible transcription errors
- Added one line to make YT summary task explicitly aware of possible transcript typos.

Changes
PR #2046 by ksylvan: refactor: add GetRaw method to encapsulate raw pattern loading logic
- Add GetRaw method to PatternsEntity for unprocessed pattern retrieval
- Replace inline raw pattern loading logic in server handler with GetRaw
- Remove manual Pattern struct construction from PatternsHandler.Get
- Simplify server handler by delegating storage access to database layer
- Add test coverage for GetRaw with custom patterns directory

Changes
PR #2040 by Twozee-Tech: Add Polish (pl) locale
- Add Polish (pl) locale with a complete translation file containing all 683 keys, matching the structure of existing locale files in internal/i18n/locales/.

Changes
PR #2038 by ksylvan: Add Git CLI Fallback to FetchFilesFromRepo
- Feat: add git CLI fallback to FetchFilesFromRepo for improved compatibility
- Add git CLI fallback when go-git in-memory clone fails, with detection via exec.LookPath
- Implement fetchFilesViaGitCLI using a shallow --depth 1 clone into a temp directory with deferred cleanup
- Respect SingleDirectory and PathPrefix options in the CLI path, and surface a combined error when both go-git and CLI fallback fail
- Extract fetchFilesViaGoGit into a dedicated helper function and add a copyFile helper to support CLI-based file extraction

Changes
PR #2036 by ksylvan: WebUI: YouTube URL Processing Refactor & Dependency Updates
- Feat: Inline YouTube transcripts into the chat flow, replacing YouTube URLs with fetched transcript blocks and supporting multiple links per message
- Feat: Always render the user message before processing the response, with a YouTube-specific loading indicator displayed during transcript fetch
- Fix: Reset YouTube URL state after submission processing to prevent stale data across messages
- Chore: Upgrade SvelteKit, Svelte, Tailwind, and TypeScript toolchain versions, along with bumped pdfjs-dist and @napi-rs/canvas optional dependency versions
- Chore: Regenerate pnpm/npm lockfiles to reflect updated transitive dependencies

Changes
PR #2033 by dependabot and ksylvan: Web UI: Dependabot upgrades and bug fixes
- Update Svelte to version 5.53.5 and upgrade @sveltejs/vite-plugin-svelte to 4.0.0, keeping core dependencies current
- Upgrade Rollup to version 4.59.0 and lucide-svelte to version 0.575.0 for improved bundling and iconography support
- Replace custom npm and pnpm install scripts with a standardized postinstall script for svelte-kit sync, simplifying the setup process
- Fix self-closing tags in Svelte components and update transcript joining to use newlines, resolving rendering and formatting issues
- Make the cleanPatternOutput method public in ChatService and remove the svelte-markdown dependency, reducing bundle size and improving API accessibility
PR #2034 by konstantint and ksylvan: feat: add create_slides pattern
- Add new create_slides pattern for generating Reveal.js HTML slideshows
- Register create_slides across CONVERSION, VISUALIZE, and WRITING categories
- Add create_slides to the pattern explanations index and suggest_pattern user guide
- Insert create_slides entry into pattern_descriptions.json and pattern_extracts.json with relevant tags
- Shift pattern numbering from 97 onward to accommodate the new entry

Changes
PR #2025 by [dependabot]](https://github.com/apps/dependabot): chore(deps): bump the npm_and_yarn group across 1 directory with 5 updates
- Updated @sveltejs/kit from version 2.49.5 to 2.53.1 in the /web directory.
- Updated svelte from version 4.2.20 to 5.51.5 in the /web directory.
- Updated ajv from version 6.12.6 to 6.14.0 in the /web directory.
- Updated devalue from version 5.6.2 to 5.6.3 in the /web directory.
- Updated minimatch from version 3.1.2 to 3.1.4 in the /web directory.

Changes
PR #2032 by ksylvan: Refactor: Consolidate NeedsRawMode Default Implementation into PluginBase
- Refactor: move NeedsRawMode default implementation to PluginBase
- Add default NeedsRawMode method to shared PluginBase struct
- Remove redundant NeedsRawMode implementations across all AI vendor plugins
- Consolidate default false return logic into single base plugin method
- Clean up duplicate boilerplate from anthropic, bedrock, copilot, gemini plugins

Changes
PR #2031 by ksylvan: Web UI: Add Vendor Filter Dropdown to Model Selection UI
- Add a vendor selector dropdown to filter available models in the chat UI
- Introduce selectedVendor writable store and vendorNames derived store with sorted, unique vendor names
- Add filteredModels derived store to dynamically filter models by the selected vendor
- Fix model deduplication to use a vendor:name composite key, with models sorted by vendor then name (case-insensitive)
- Display the vendor prefix in model option labels for improved clarity

Changes
PR #2029 by ksylvan: Web UI: Fix streaming chat: tokens now accumulate in order into a single message
- Feat: improve streaming message handling and SSE buffer parsing
- Append content to existing assistant messages instead of replacing, and fix loading message removal to search by index rather than position
- Refactor SSE buffer splitting to always retain incomplete segments, and trim segments before parsing to handle whitespace edge cases
- Consolidate duplicate message update logic across chat components and process remaining buffer after stream completion more reliably

Changes
PR #2021 by jlec and ksylvan: Enhanced Azure AI Gateway with i18n Support and documentation
- Added configurable API version support for the Azure OpenAI backend, defaulting to 2025-04-01-preview while maintaining backward compatibility.
- Implemented URL validation with HTTPS enforcement and a cancellable context in SendStream with a 300-second timeout.
- Added a 10MB response body size limit using io.LimitReader and improved error body truncation from 200 to 500 characters.
- Added file-level documentation across all backend files and enforced lowercase error messages per Go convention.
- Fixed Bedrock max_tokens to correctly respect opts.MaxTokens with a fallback, and added error checks for empty message lists across all backends.

Changes
PR #2020 by ksylvan: Add wire Debug Level for LLM Request/Response Logging
- Feat: add wire debug level (4) for full LLM request/response debug logging
- Add Wire log level constant to debug level enum and expose GetLevel() for safe concurrent level reads
- Log outbound message roles/content, inbound stream updates, token usage, and non-streaming LLM responses at wire debug level
- Update --debug flag description and set_debug_level locale strings across all 10 languages to include new level 4
- Update zsh, bash, and fish shell completions to include the new 4 (wire) debug level value

Changes
PR #2019 by ksylvan: feat: replace hardcoded error strings with i18n translation keys
- Replace hardcoded error messages with i18n.T() calls across Go source files, covering chat, attachment, storage, template, plugin registry, server, and utility packages
- Added approximately 80 new translation keys to all locale files (en, de, es, fa, fr, it, ja, pt-BR, pt-PT, zh), with keys sorted alphabetically
- Internationalized error strings across githelper, notifications, patterns, sessions, and provider modules (DigitalOcean, Gemini, OpenAI-compatible)
- Refactored fmt.Errorf("%s", ...) usages to errors.New() and normalized i18n error strings to lowercase per Go conventions
- Updated db_error_loading_env_file format verb from %s to %w to support proper error wrapping

Changes
PR #2014 by jlec and ksylvan: feat: add Azure AI Gateway plugin support
- Added Azure AI Gateway plugin to the plugin registry, enabling it as an available AI provider
- Imported the new azureaigateway plugin package and registered azureaigateway.NewClient() in the plugin registry

Changes
PR #2013 by ByronPogson and ksylvan: feat: add Azure Entra ID authentication plugin
- Added a new Azure Entra ID authentication plugin with shared Azure utilities, integrating it into the plugin registry.
- Extracted shared Azure logic into a new azurecommon package, consolidating ParseDeployments, BuildEndpoint, and middleware.
- Upgraded azidentity to v1.13.1 with Entra ID/MSAL support, and added golang-jwt, pkg/browser, and go-keychain as new dependencies.
- Added azure_credential_failure and azure_base_url_question i18n keys across all locales, and enforced non-empty validation for Azure deployment names on configure.

Changes
PR #2016 by ksylvan: Extend Anthropic model beta map with 1M context models
- Extends the Anthropic model beta map to include 1M context window models, adding Claude Sonnet 4.6, Claude Opus 4.5, and Claude Opus 4.6 to the beta entries.
- Documents 1M token context window model support and clarifies the model beta list maintenance and update strategy.
- Groups model variants under clearer, annotated sections for improved readability and organization.

Changes
PR #2012 by ksylvan: Remove unused gemini_openai plugin and oauth_storage utility
- Removed the unused gemini_openai plugin package, including its OpenAI-compatible client wrapper.
- Removed the OAuthToken struct and its expiry-check logic from the util package.
- Removed the OAuthStorage utility, including persistent token save, load, and delete handlers.
- Dropped the HasValidToken helper function and atomic file-write token saving mechanism.
- Deleted all oauth_storage unit tests covering the token lifecycle.

Changes
PR #2011 by ksylvan: Add support for Claude Sonnet 4.6
- Upgraded the anthropic-sdk-go dependency from v1.22.0 to v1.23.0 to support the latest Anthropic SDK features.
- Added Claude Sonnet 4.6 to the list of supported models.
- Updated go.sum checksums to reflect the new SDK version.

Changes
PR #1999 by ghrom: feat: add 3 patterns from cross-model AI dialogue research
- Add audit_consent pattern for detecting manufactured consent via power asymmetry analysis, surfaced from a devil's advocate "consent theater" critique across multi-model stress testing.
- Add detect_silent_victims pattern to identify harmed parties who cannot speak for themselves, including future generations and unaware victims.
- Add audit_transparency pattern for evaluating whether decisions are explainable to affected parties across five dimensions.
- Register all three new patterns in pattern descriptions and extracts JSON files, and categorize them under ANALYSIS and CR THINKING in suggest_pattern.
- Update pattern_explanations.md with renumbered entries to reflect the newly added patterns.
Direct commits
- Fix: update Buy Me a Coffee links to correct profile URL

Changes
PR #2006 by konstantint: feat: When running from a symlink, use the executable name as the pattern argument
- Feat: When running from a symlink, use the executable name as the pattern argument

Changes
PR #2007 by ksylvan: Add optional API key authentication to LM Studio client
- Add optional API key authentication to LM Studio client
- Add optional API key setup question to client configuration
- Add ApiKey field to the LM Studio Client struct
- Create addAuthorizationHeader helper to attach Bearer token to requests
- Apply authorization header to all outgoing HTTP requests

Changes
PR #2005 by ksylvan: I18N: For file manager, Vertex AI, and Copilot errors
- Internationalized file manager, Vertex AI, and Copilot error messages via i18n by replacing hardcoded error strings with translation keys
- Added file manager, Vertex AI, and Copilot i18n keys to all 10 locale files
- Fixed JSON trailing comma syntax errors across all locale files
- Normalized German locale JSON indentation from tabs to spaces
- Updated Bedrock AWS region setup to use AddSetupQuestionWithEnvName

Changes
PR #2004 by ksylvan: Add i18n translations for VertexAI, Gemini, Bedrock, and fetch plugins
- Add i18n translations for VertexAI, Gemini, Bedrock, and fetch plugins across 10 locale files
- Replace hardcoded English strings with i18n.T() calls in Bedrock, Gemini, VertexAI, and fetch plugins
- Add error handling for fetch operations with new error messages in i18n
- Use errors.New instead of fmt.Errorf for non-formatted error strings
- Add Gemini TTS and audio error translations, AWS Bedrock client error translations, and fetch plugin error translations to all locales

Changes
PR #2002 by ksylvan: Internationalization Polish
- Added i18n translations for ollama, extensions, lmstudio, and spotify modules
- Replaced hardcoded English strings with i18n.T() calls across all modules
- Added translations for all new keys in de, en, es, fa, fr, it, ja, pt-BR, pt-PT, and zh locales
- Added i18n strings for extension registry, executor, and manager operations
- Added i18n strings for Spotify API client error messages
PR #2003 by ksylvan: Add internationalization support for chatter and template file operations
- Add internationalization support for chatter and template file operations
- Replace hardcoded strings with i18n keys in chatter.go and file.go
- Provide translations in nine languages: German, English, Spanish, Persian, French, Italian, Japanese, Portuguese, and Chinese
- Enable localized output for stream updates and file plugin operations
- Maintain backward compatibility with existing functionality
Direct commits
- MAESTRO: i18n: extract hard-coded strings from internal/tools/spotify/spotify.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/plugins/template/extension_executor.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/plugins/ai/openai/openai.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/plugins/template/extension_registry.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/plugins/ai/lmstudio/lmstudio.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/plugins/template/extension_manager.go
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com
- MAESTRO: i18n: extract hard-coded strings from internal/server/ollama.go
Replace 37 hard-coded error/log strings with i18n.T() calls and add
translations for all 10 supported languages (en, de, es, fa, fr, it,
ja, pt-BR, pt-PT, zh). Keys use ollama_ prefix following project
conventions.
Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com

Changes
PR #1997 by ksylvan: Enable responses API usage for GrokAI (xAI) provider
- Enable responses API for GrokAI (xAI) provider
- Set ImplementsResponses to true for GrokAI provider

Changes
PR #1993 by ksylvan: Add dynamic Abacus RouteLLM model list and update fallback static list
- Add dynamic model fetching from Abacus RouteLLM API with fetchAbacusModels() function and fallback to static list on failure
- Add GPT-5 Codex variants (5, 5.1, 5.1-max, 5.2) and GPT-5.2-chat-latest to static model list
- Add claude-opus-4-6, gemini-3-flash-preview, kimi-k2.5, glm-4.7, and glm-5 to static models
- Rename qwen/qwen3-Max to lowercase qwen3-max for consistency
- Handle static:abacus as special case in ListModels routing logic

Changes
Release Notes
PR #1988 by ghrom and ksylvan: feat: add Ultimate Law AGI safety pattern suite
- Add four patterns implementing minimal, falsifiable ethical constraints for AGI safety evaluation
- Add ultimate_law_safety pattern for evaluating actions against "no unwilling victims" principle
- Add detect_mind_virus pattern for identifying manipulative reasoning that resists correction
- Add check_falsifiability pattern for verifying claims can be tested and proven wrong
- Add extract_ethical_framework pattern for surfacing implicit ethics in documents and policies
PR #1990 by davejpeters and ksylvan: add pattern explain_terms_and_conditions
- Added explain_terms_and_conditions pattern for comprehensive legal agreement analysis with focus on consumer protection
- Renamed suggest_moltbot_command to suggest_openclaw_pattern with updated branding
- Registered new patterns in pattern descriptions and extracts JSON
- Updated suggest_pattern documentation with new pattern summaries
- Removed deprecated suggest_moltbot_command pattern
PR #1991 by ksylvan: Upgrade project dependencies including AWS, Anthropic, and Google SDKs
- Chore: upgrade project dependencies including AWS, Anthropic, and Google SDKs
- Update Anthropic SDK to version 1.22.0 for improved functionality
- Bump AWS SDK and Bedrock runtime to improve cloud integration
- Upgrade Ollama dependency to version 0.15.6 for model support
- Refresh Google API and GenAI libraries to current stable releases

Changes
PR #1986 by ksylvan: Support Anthropic Opus 4.6
- Upgrade anthropic-sdk-go from v1.20.0 to v1.21.0
- Add ClaudeOpus4_6 to supported Anthropic model list
- Remove unused indirect dependencies from go.mod and go.sum
- Clean up legacy protobuf and gRPC version references
- Drop unused table writer and console dependencies

Changes
PR #1983 by dependabot: chore(deps): bump @isaacs/brace-expansion from 5.0.0 to 5.0.1 in /web in the npm_and_yarn group across 1 directory
- Updated @isaacs/brace-expansion dependency from version 5.0.0 to 5.0.1 in the web directory

Changes
PR #1979 by ksylvan: Update Anthropic SDK to v1.20.0 and reorganize model definitions
- Feat: update Anthropic SDK to v1.20.0 and reorganize model definitions
- Bump anthropic-sdk-go dependency from v1.19.0 to v1.20.0
- Add deprecation notice for pre-February 2026 legacy models
- Add new Claude Sonnet 4.0 and Opus 4.0 model aliases
- Extend 1M context beta support to all Sonnet 4 variants

Changes
PR #1975 by koriyoshi2041: feat: add suggest_moltbot_command pattern for Moltbot (formerly Clawdbot) CLI
- Added new pattern for suggesting Moltbot CLI commands based on natural language intent
- Fixed multi-command output format inconsistency to preserve pipe-friendly behavior
- Updated all CLI references and command examples to use new moltbot binary name
- Added new dictionary words for VSCode spellcheck and fixed markdown table formatting
PR #1978 by ksylvan: chore: remove OAuth support from Anthropic client
- Remove OAuth support from Anthropic client and delete related OAuth files
- Simplify configuration handling to check only API key instead of OAuth credentials
- Clean up imports and unused variables in anthropic.go
- Update server configuration methods to remove OAuth references
- Remove OAuth-related environment variables from configuration
Direct commits
- Docs: fix ChangeLog snippet for PR 1975

Changes
PR #1972 by ksylvan: More node package updates: remove cn, fix string and request vulnerabilities
- Removed cn (Chuck Norris jokes) package to resolve security vulnerabilities
- Fixed 5 Dependabot alerts including ReDoS vulnerabilities in string package and SSRF/Remote Memory Exposure issues in request package
- Enhanced security posture by eliminating vulnerable dependencies with no available patches

Changes
PR #1968 by ksylvan: New extract_all_quotes and move misplaced patterns
- Move orphaned pattern files from patterns/ to data/patterns/
- Add extract_all_quotes pattern for quote extraction
- The extract_all_quotes is originally from PR #176 by @CPAtoCybersecurity
- The suggest_pattern pattern is updated with the following additions.
- extract_bd_ideas for actionable idea extraction
- suggest_gt_command for GT command suggestions
- create_bd_issue for issue tracking commands

Changes
MiniMax unexpected status code: 404 from provider MiniMax, response body: 404 page not found
PR #1965 by infinitelyloopy-bt: fix(azure): Fix deployment URL path for Azure OpenAI API
- Fixed deployment URL path construction for Azure OpenAI API to correctly include deployment names in request URLs
- Added custom middleware to transform API paths and extract deployment names from request body model fields
- Moved StreamOptions configuration to only apply to streaming requests, as Azure rejects stream_options for non-streaming requests
- Added Azure OpenAI troubleshooting documentation with technical details and configuration guidance
- Resolved SDK route matching bug that was preventing proper URL generation for Azure OpenAI endpoints

Changes
MiniMax unexpected status code: 404 from provider MiniMax, response body: 404 page not found
PR #1964 by jessesep: feat: add design system, golden rules, and discord structure patterns
- Added create_design_system pattern to generate CSS design systems with tokens, typography scales, and dark/light mode support from requirements
- Added create_golden_rules pattern to extract implicit and explicit rules from codebases into testable, enforceable guidelines
- Added analyze_discord_structure pattern to audit Discord server organization, permissions, and naming conventions
PR #1967 by ksylvan: chore: add MiniMax provider support and update API endpoints
- Add MiniMax provider configuration with API endpoint updates
- Implement NeedsRawMode method for MiniMax model handling
- Define static MiniMax model list with M2 variants
- Add Infermatic and Novita to VS Code extensions
- Configure MiniMax models as static discovery

Changes
PR #1960 by ksylvan: fix: consume all positional arguments as input
- Fix: consume all positional arguments as input by joining all positional arguments with spaces instead of using only the last argument, allowing commands to process entire phrases correctly

Changes
PR #1957 by ksylvan: Add Novita AI as a new OpenAI-compatible provider
- Add Novita AI as a new OpenAI-compatible provider
- Add Novita AI provider configuration with API endpoint
- Update README to include Novita AI in supported providers list
- Configure Novita AI to use OpenAI-compatible interface

Changes
PR #1950 by cleong14: feat: add suggest_gt_command pattern
- Added new suggest_gt_command pattern that suggests Gas Town (gt) CLI commands based on natural language descriptions
- Covers 85+ commands across work management, agents, communication, services, diagnostics, and recovery
- Includes pipe-friendly output with raw command on first line for easy command extraction
- Integrates with Gas Town CLI tool for enhanced command-line workflow automation
PR #1951 by ksylvan: Add extract_wisdom_with_attribution pattern for speaker-attributed quotes
- Add new extract_wisdom_with_attribution pattern that extends extract_wisdom functionality with speaker attribution capabilities
- Create comprehensive documentation including README and system.md files for the new pattern
- Update pattern_explanations.md with detailed entry for the new pattern
- Add pattern to suggest_pattern category lists for improved discoverability
- Update pattern metadata files including pattern_descriptions.json and pattern_extracts.json with relevant tags and content
PR #1952 by ksylvan: Fix: using attachments with Anthropic models
- Feat: add multi-content support for images and PDFs in Anthropic client
- Update toMessages to handle multi-content messages with text and attachments
- Add contentBlocksFromMessage to convert message parts to Anthropic blocks
- Implement support for image URLs including data URLs and base64 images
- Add PDF attachment handling via data URLs and URL-based PDFs

Changes
PR #1945 by ksylvan: feat: Add Spotify API integration for podcast metadata retrieval
- Add Spotify metadata retrieval via --spotify flag
- Add Spotify plugin with OAuth token handling and metadata
- Wire --spotify flag into CLI processing and output
- Register Spotify in plugin setup, env, and registry
- Update shell completions to include --spotify option

Changes
PR #1947 by cleong14: feat(patterns): add extract_bd_ideas pattern
- Added extract_bd_ideas pattern that extracts actionable ideas from content and transforms them into well-structured bd issue tracker commands
- Implemented identification system for tasks, problems, ideas, improvements, bugs, and features
- Added actionability evaluation and appropriate scoping functionality
- Integrated priority assignment system (P0-P4) with relevant labels
- Created ready-to-execute bd create commands output format
PR #1948 by cleong14: feat(patterns): add create_bd_issue pattern
- Added create_bd_issue pattern that transforms natural language issue descriptions into optimal bd (Beads) issue tracker commands
- Implemented comprehensive bd create flag reference for better command generation
- Added intelligent type detection system that automatically categorizes issues as bug, feature, task, epic, or chore
- Included priority assessment capability that assigns P0-P4 priority levels based on urgency signals in descriptions
- Integrated smart label selection feature that automatically chooses 1-4 relevant labels for each issue
PR #1949 by ksylvan: Fix #1931 - Image Generation Feature should warn if the model is not capable of Image Generation
- Add image generation compatibility warnings for unsupported models
- Add warning to stderr when using incompatible models with image generation
- Add GPT-5, GPT-5-nano, and GPT-5.2 to supported image generation models
- Create checkImageGenerationCompatibility function in OpenAI plugin
- Add comprehensive tests for image generation compatibility warnings

Changes
PR #1944 by ksylvan: Add Infermatic AI Provider Support
- Add Infermatic provider to ProviderMap as part of Phase 1 implementation for issue #1033
- Add test coverage for the Infermatic AI provider in TestCreateClient to verify provider exists and creates valid client
- Replace go-git status API with native git status --porcelain command to fix worktree compatibility issues
- Simplify IsWorkingDirectoryClean and GetStatusDetails functions to use CLI output parsing instead of go-git library
- Use native git rev-parse HEAD to get commit hash after commit and remove unused imports from walker.go

Changes
PR #1943 by ksylvan: fix: Ollama server now respects the default context window
- Fix: Ollama server now respects the default context window instead of using hardcoded 2048 tokens
- Add parseOllamaNumCtx() function with type-safe extraction supporting 6 numeric types and platform-aware integer overflow protection
- Extract num_ctx from client request options and add ModelContextLength field to ChatRequest struct
- Implement DoS protection via 1,000,000 token maximum limit with sanitized error messages
- Add comprehensive unit tests for parseOllamaNumCtx function covering edge cases including overflow and invalid types

Changes
PR #1941 by ksylvan: Add greybeard_secure_prompt_engineer to metadata, also remove duplicate json data file
- Add greybeard_secure_prompt_engineer pattern to metadata (pattern explanations and json index)
- Refactor build process to use npm hooks for copying JSON files instead of manual copying
- Update .gitignore to exclude generated data and tmp directories
- Modify suggest_pattern categories to include new security pattern
- Delete redundant web static data file and rely on build hooks

Changes
PR #1940 by ksylvan: Rewrite Ollama chat handler to support proper streaming responses
- Refactor Ollama chat handler to support proper streaming responses with real-time SSE data parsing
- Replace single-read body parsing with streaming bufio.Scanner approach and implement writeOllamaResponse helper function
- Add comprehensive error handling improvements including proper HTTP error responses instead of log.Fatal to prevent server crashes
- Fix upstream error handling to return stringified error payloads and validate Fabric chat URL hosts
- Implement proper request context propagation and align duration fields to int64 nanosecond precision for consistency

Changes
PR #1936 by ksylvan: New Vendor: Microsoft Copilot
- Add Microsoft 365 Copilot integration as a new AI vendor with OAuth2 authentication for delegated user permissions
- Enable querying of Microsoft 365 data including emails, documents, and chats with both synchronous and streaming response support
- Provide comprehensive setup instructions for Azure AD app registration and detail licensing, technical, and permission requirements
- Add troubleshooting steps for common authentication and API errors with current API limitations documentation
- Fix SendStream interface to use domain.StreamUpdate instead of chan string to match current Vendor interface requirements

Changes
PR #1933 by ksylvan: Add DigitalOcean Gradient AI support
- Feat: add DigitalOcean Gradient AI Agents as a new vendor
- Add DigitalOcean as a new AI provider in plugin registry
- Implement DigitalOcean client with OpenAI-compatible inference endpoint
- Support model access key authentication for inference requests
- Add optional control plane token for model discovery
Direct commits
- Chore: Update README with a links to other docs

Changes
PR #1929 by ksylvan: Add Mammouth as new OpenAI-compatible AI provider
- Feat: add Mammouth as new OpenAI-compatible AI provider
- Add Mammouth provider configuration with API base URL
- Configure Mammouth to use standard OpenAI-compatible interface
- Disable Responses API implementation for Mammouth provider
- Add "Mammouth" to VSCode spell check dictionary

Changes
PR #1928 by ksylvan: Eliminate repetitive boilerplate across eight vendor implementations
- Refactor: add NewVendorPluginBase factory function to reduce duplication
- Update 8 vendor files (anthropic, bedrock, gemini, lmstudio, ollama, openai, perplexity, vertexai) to use the factory function
- Add 3 test cases for the new factory function
- Add centralized factory function for AI vendor plugin initialization
- Chore: exempt json files from VSCode format-on-save
Direct commits
- Docs: Add GitHub sponsor section to README
I spend hundreds of hours a year on open source. If you'd like to help support this project, you can sponsor me here.
Co-Authored-By: Claude Opus 4.5 noreply@anthropic.com

Changes
PR #1925 by ksylvan: docs: update README to document new AI providers and features
- Docs: update README to document new AI providers and features
- List supported native and OpenAI-compatible AI provider integrations
- Document dry run mode for previewing prompt construction
- Explain Ollama compatibility mode for exposing API endpoints
- Detail available prompt strategies like chain-of-thought and reflexion
PR #1926 by henricook and ksylvan: feat(vertexai): add dynamic model listing and multi-model support
- Dynamic model listing from Vertex AI Model Garden API
- Support for both Gemini (genai SDK) and Claude (Anthropic SDK) models
- Curated Gemini model list with web search support for Gemini models
- Thinking/extended thinking support for Gemini
- TopP parameter support for Claude models

Changes
PR #1914 by majiayu000: feat(code_helper): add stdin support for piping file lists
- Added stdin support for piping file lists to code_helper, enabling commands like find . -name '*.go' | code_helper "instructions" and git ls-files '*.py' | code_helper "Add type hints"
- Implemented automatic detection of stdin pipe mode with single argument (instructions) support
- Enhanced tool to read file paths from stdin line by line while maintaining backward compatibility with existing directory scanning functionality
PR #1915 by majiayu000: feat: parallelize audio chunk transcription for improved performance
- Parallelize audio chunk transcription using goroutines for improved performance

Changes
PR #1921 by ksylvan: chore: remove redundant --sync-db step from changelog workflow
- Remove redundant --sync-db step from changelog workflow
- Remove duplicate database sync command from version workflow
- Simplify changelog generation to single process-prs step
- Clean up heal_person pattern by removing duplicate content sections
- Remove duplicate IDENTITY, PURPOSE, STEPS, and OUTPUT INSTRUCTIONS from pattern file

Changes
PR #1919 by ksylvan: Fix the last_pr_sync setting during PR incoming processing
- Fix: update SetLastPRSync to use version date instead of current time
- Change last_pr_sync to use versionDate instead of time.Now()
- Ensure future runs fetch PRs merged after the version date
- Add clarifying comments explaining the sync timing logic

Changes
PR #1912 by berniegreen: refactor: implement structured streaming and metadata support
- Feat: add domain types for structured streaming (Phase 1)
- Refactor: update Vendor interface and Chatter for structured streaming (Phase 2)
- Refactor: implement structured streaming in all AI vendors (Phase 3)
- Feat: implement CLI support for metadata display (Phase 4)
- Feat: implement REST API support for metadata streaming (Phase 5)

Changes
PR #1907 by majiayu000: feat(gui): add Session Name support for multi-turn conversations
- Added Session Name support for multi-turn conversations in GUI chat interface, enabling persistent conversations similar to CLI's --session flag
- Added SessionName field to PromptRequest and sessionName to ChatPrompt interface for proper session handling
- Extracted SessionSelector component with Select component instead of native dropdown for better user experience
- Implemented session message loading when selecting existing sessions with proper error handling
- Fixed Select component binding and empty input handling to prevent redundant API calls and properly clear sessions

Changes
PR #1906 by ksylvan: Code Quality: Optimize HTTP client reuse + simplify error formatting
- Refactor: optimize HTTP client reuse and simplify error formatting
- Simplify error wrapping by removing redundant Sprintf calls in CLI
- Pass HTTP client to FetchModelsDirectly to enable connection reuse
- Store persistent HTTP client instance inside the OpenAI provider struct
- Update compatible AI providers to match the new function signature

Changes
PR #1904 by majiayu000: fix: resolve WebUI tooltips not rendering due to overflow clipping
- Fix: resolve WebUI tooltips not rendering due to overflow clipping by using position: fixed and getBoundingClientRect() to calculate tooltip position dynamically, preventing tooltips from being clipped by parent containers with overflow: hidden
- Refactor: extract tooltip positioning logic into separate positioning.ts module for better code organization and maintainability
- Improve accessibility with aria-describedby attributes and unique IDs for better screen reader support
- Add reactive tooltip position updates on scroll and resize events for dynamic positioning
- Add SSR safety with isBrowser flag check and comprehensive unit test coverage for the positioning functions

Changes
PR #1890 by ksylvan: Bundle yt-dlp with fabric in Nix flake, introduce slim variant
- Added yt-dlp bundling with fabric package and introduced fabric-slim variant
- Renamed original fabric package to fabricSlim and created new fabric package as symlinkJoin of fabricSlim and yt-dlp
- Added fabric-slim output for the slim variant and updated default package to point to bundled fabric
- Enhanced fabric meta description to note yt-dlp inclusion and set mainProgram to fabric in bundled package
- Added wrapper for fabric binary to include PATH in execution environment

Changes
PR #1887 by bvandevliet: feat: correct video title and added description to yt transcript api response
- Feat: correct video title (instead of id) and added description to yt transcript api response
- Updated API documentation.

Changes
PR #1877 by ksylvan: modernize: update GitHub Actions and modernize Go code
- Modernize GitHub Actions and Go code with latest stdlib features
- Upgrade GitHub Actions to latest versions (v6, v21) and add modernization check step
- Replace strings manipulation with strings.CutPrefix and strings.CutSuffix
- Replace manual loops with slices.Contains for validation and use strings.SplitSeq for iterator-based splitting
- Replace fmt.Sprintf with fmt.Appendf for efficiency and simplify padding calculation with max builtin

Changes
PR #1867 by jaredmontoya: chore: update flake
- Chore: update flake
- Merge branch 'main' into update-flake

Changes
PR #1856 by ksylvan: Add support for new ClaudeHaiku 4.5 models
- Add support for new ClaudeHaiku models in client
- Add ModelClaudeHaiku4_5 to supported models
- Add ModelClaudeHaiku4_5_20251001 to supported models

Changes
PR #1852 by ksylvan: Add Abacus vendor for ChatLLM models with static model list
- Add static model support and register Abacus provider
- Detect modelsURL starting with 'static:' and route appropriately
- Implement getStaticModels returning curated Abacus model list
- Register Abacus provider with ModelsURL 'static:abacus'
- Extend provider tests to include Abacus existence

Changes
PR #1851 by ksylvan: Add Z AI provider and glm model support
- Add Z AI provider configuration to ProviderMap
- Include BaseURL for Z AI API endpoint
- Add test case for Z AI provider existence
- Add glm to OpenAI model prefixes list
- Support new Z AI provider in OpenAI compatible plugins

Changes
PR #1847 by ksylvan: Improve model name matching for NeedsRaw in Ollama plugin
- Improved model name matching in Ollama plugin by replacing prefix-based matching with substring matching
- Enhanced NeedsRaw functionality to support more flexible model name detection
- Renamed ollamaPrefixes variable to ollamaSearchStrings for better code clarity
- Replaced HasPrefix function with Contains for more comprehensive model matching
- Added "conceptmap" to VSCode dictionary settings
Direct commits
- Merge branch 'danielmiessler:main' into main
- Docs: Fix typo in README

Changes
PR #1845 by ksylvan: Add Claude Opus 4.5 Support
- Add Claude Opus 4.5 model variants to Anthropic client
- Upgrade anthropic-sdk-go from v1.16.0 to v1.19.0
- Update golang.org/x/crypto from v0.41.0 to v0.45.0
- Upgrade golang.org/x/net from v0.43.0 to v0.47.0
- Bump golang.org/x/text from v0.28.0 to v0.31.0

Changes
PR #1843 by ksylvan: Implement case-insensitive vendor and model name matching
- Fix: implement case-insensitive vendor and model name matching across the application
- Add case-insensitive vendor lookup in VendorsManager
- Implement model name normalization in GetChatter method
- Add FilterByVendor method with case-insensitive matching
- Add FindModelNameCaseInsensitive helper for model queries

Changes
PR #1839 by ksylvan: Add GitHub Models Provider and Refactor Fetching Fallback Logic
- Add GitHub Models provider and refactor model fetching with direct API fallback
- Add GitHub Models to supported OpenAI-compatible providers list
- Implement direct HTTP fallback for non-standard model responses
- Centralize model fetching logic in openai package
- Upgrade openai-go SDK dependency from v1.8.2 to v1.12.0

Changes
PR #1836 by ksylvan: docs: clarify --raw flag behavior for OpenAI and Anthropic providers
- Update --raw flag description across all documentation files
- Clarify flag only affects OpenAI-compatible providers behavior
- Document Anthropic models use smart parameter selection
- Remove outdated reference to system/user role changes
- Update help text in CLI flags definition

Changes
PR #1831 by ksylvan: Remove get_youtube_rss pattern
- Chore: remove get_youtube_rss pattern from multiple files
- Remove get_youtube_rss from pattern_explanations.md
- Delete get_youtube_rss entry in pattern_descriptions.json
- Delete get_youtube_rss entry in pattern_extracts.json
- Remove get_youtube_rss from suggest_pattern/system.md
PR #1832 by ksylvan: Improve channel management in Gemini provider
- Fix: improve channel management in Gemini streaming method
- Add deferred channel close at function start
- Return error immediately instead of breaking loop
- Remove redundant channel close statements from loop
- Ensure channel closes on all exit paths consistently

Changes
PR #1802 by nickarino: fix: improve template extension handling for {{input}} and add examples
- Fix: improve template extension handling for {{input}} and add examples
PR #1823 by ksylvan: Add missing patterns and renumber pattern explanations list
- Add apply_ul_tags pattern for content categorization
- Add extract_mcp_servers pattern for MCP server identification
- Add generate_code_rules pattern for AI coding guardrails
- Add t_check_dunning_kruger pattern for competence assessment
- Renumber all patterns from 37-226 to 37-230
Direct commits
- Chore: incoming 1823 changelog entry

Changes
PR #1814 by ksylvan: Add Concept Map in html
- Add create_conceptmap for interactive HTML concept maps using Vis.js
- Add fix_typos for proofreading and correcting text errors
- Introduce model_as_sherlock_freud for psychological modeling and behavior analysis
- Implement predict_person_actions for behavioral response predictions
- Add recommend_yoga_practice for personalized yoga guidance
- Credit goes to @FELIPEGUEDESBR for the pattern
PR #1816 by ksylvan: Update anthropic-sdk-go to v1.16.0 and update models
- Upgraded anthropic-sdk-go from v1.13.0 to v1.16.0
- Removed outdated model ModelClaude3_5SonnetLatest
- Added new model ModelClaudeSonnet4_5_20250929
- Updated anthropic beta map to include the new model
- Updated dependencies in go.sum file

Changes
PR #1803 by dependabot[bot][bot]: chore(deps-dev): bump vite from 5.4.20 to 5.4.21 in /web in the npm_and_yarn group across 1 directory
- Updated Vite development dependency from version 5.4.20 to 5.4.21 in the web directory
PR #1805 by OmriH-Elister: Added several new patterns
- Added new WELLNESS category with four patterns including personalized yoga practice recommendations and wellness guidance
- Added model_as_sherlock_freud pattern for psychological detective analysis combining Sherlock Holmes deduction with Freudian psychology
- Added predict_person_actions pattern for behavioral response predictions based on personality analysis
- Added fix_typos pattern for automated proofreading and typo corrections
- Updated ANALYSIS and SELF categories to include new wellness-related patterns and classifications
PR #1808 by sluosapher: Updated create_newsletter_entry pattern to generate more factual titles
- Updated the title generation style; added an output example.

Changes
PR #1780 by marcas756: feat: add extract_characters pattern
- Define character extraction goals and steps with canonical naming and deduplication rules
- Outline interaction mapping and narrative importance analysis
- Provide comprehensive output schema with proper formatting guidelines
- Include positive and negative examples for pattern clarity
- Enforce restrictions on speculative motivations and non-actor inclusion
PR #1794 by starfish456: Enhance web app docs
- Remove duplicate content from the main readme and link to the web app readme
- Update table of contents with proper nesting and fix minor formatting issues
PR #1810 by tonymet: improve subtitle lang, retry, debugging & error handling
- Improve subtitle lang, retry, debugging & error handling
Direct commits
- Docs: clean up README - remove duplicate image and add collapsible updates section
- Remove duplicate fabric-summarize.png screenshot
- Wrap Updates section in HTML details/summary accordion to save space
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude noreply@anthropic.com
- Updated CSE pattern.

Changes
PR #1783 by ksylvan: Update anthropic-sdk-go and add claude-sonnet-4-5
- Feat: update anthropic-sdk-go to v1.13.0 and add new model
- Upgrade anthropic-sdk-go to version 1.13.0
- Add ModelClaudeSonnet4_5 to supported models list

Changes
PR #1779 by ksylvan: Improve pt-BR Translation - Thanks to @JuracyAmerico
- Fix: improve PT-BR translation naturalness and fluency
- Replace "dos" with "entre" for better preposition usage
- Add definite articles where natural in Portuguese
- Clarify "configurações padrão" instead of just "padrões"
- Keep technical terms visible like "padrões/patterns"

Changes
PR #1778 by ksylvan: Add Portuguese Language Variants Support (pt-BR and pt-PT)
- Add Brazilian Portuguese (pt-BR) translation file
- Add European Portuguese (pt-PT) translation file
- Implement BCP 47 locale normalization system
- Create fallback chain for language variants
- Add default variant mapping for Portuguese

Changes
PR #1777 by ksylvan: chore: remove garble installation from release workflow
- Remove garble installation step from release workflow
- Add comment for GoReleaser config file reference link
- The original idea of adding garble was to make it pass
virus scanning during version upgrades for Winget, and
this was a failed experiment.

Changes
PR #1767 by ksylvan: feat(i18n): add de, fr, ja, pt, zh, fa locales; expand tests
- Add DE, FR, JA, PT, ZH, FA i18n locale files
- Expand i18n tests with table-driven multilingual coverage
- Verify 'html_readability_error' translations across all supported languages
- Update README with release notes for added languages
- Insert blank lines between aggregated PR changelog sections
Direct commits
- Chore: update changelog formatting and sync changelog database
- Add line breaks to improve changelog readability
- Sync changelog database with latest entries
- Clean up whitespace in version sections
- Maintain consistent formatting across entries
- Chore: add spacing between changelog entries for improved readability
- Add blank lines between PR sections
- Update changelog database with to correspond with CHANGELOG fix.

Changes
PR #1759 by ksylvan: Add Windows-style Flag Support for Language Detection
- Feat: add Windows-style forward slash flag support to CLI argument parser
- Add runtime OS detection for Windows platform
- Support /flag syntax for Windows command line
- Handle Windows colon delimiter /flag:value format
- Maintain backward compatibility with Unix-style flags
PR #1762 by OmriH-Elister: New pattern for writing interaction between two characters
- Feat: add new pattern that creates story simulating interaction between two people
- Chore: add create_story_about_people_interaction pattern for persona analysis
- Add create_story_about_people_interaction pattern description
- Include pattern in ANALYSIS and WRITING categories
- Update suggest_pattern system and user documentation
Direct commits
- Chore: update alias creation to use consistent naming
- Remove redundant prefix from pattern_name variable
- Add alias_name variable for consistent alias creation
- Update alias command to use alias_name
- Modify PowerShell function to use aliasName
- Docs: add optional prefix support for fabric pattern aliases via FABRIC_ALIAS_PREFIX env var
- Add FABRIC_ALIAS_PREFIX environment variable support
- Update bash/zsh alias generation with prefix
- Update PowerShell alias generation with prefix
- Improve readability of alias setup instructions
- Enable custom prefixing for pattern commands
- Maintain backward compatibility without prefix

Changes
PR #1756 by ksylvan: Add Internationalization Support with Custom Help System
- Add comprehensive internationalization support with English and Spanish locales
- Replace hardcoded strings with i18n.T translations and add en and es JSON locale files
- Implement custom translated help system with language detection from CLI args
- Add locale download capability and localize error messages throughout codebase
- Support TTS and notification translations

Changes
PR #1755 by ksylvan: Add i18n Support for Multi-Language Fabric Experience
- Add Spanish localization support with i18n
- Create contexts and sessions tutorial documentation
- Fix broken Warp sponsorship image URL
- Remove solve_with_cot pattern from codebase
- Update pattern descriptions and explanations
Direct commits
- Update Warp sponsor section with proper formatting
- Replace with correct div structure and styling
- Use proper Warp image URL from brand assets
- Add 'Special thanks to:' text and platform availability
- Maintains proper spacing and alignment
- Fix unclosed div tag in README causing display issues
- Close the main div container properly after fabric screenshot
- Fix HTML structure that was causing repetitive content display
- Ensure proper markdown rendering on GitHub
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude noreply@anthropic.com
- Update Warp sponsor section with new banner and branding
- Replace old banner with new warp-banner-light.png image
- Update styling to use modern p tags with proper centering
- Maintain existing go.warp.dev/fabric redirect URL
- Add descriptive alt text and emphasis text for accessibility
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude noreply@anthropic.com

Changes
PR #1745 by ksylvan: Fabric Installation Improvements and Automated Release Updates
- Streamlined install process with one-line installer scripts and updated documentation
- Added bash installer script for Unix systems
- Added PowerShell installer script for Windows
- Created installer documentation with usage examples
- Simplified README installation with one-line installers

Changes
PR #1741 by ksylvan: CI: Fix Release Description Update
- Fix: update release workflow to support manual dispatch with custom tag
- Support custom tag from client payload in workflow
- Fallback to github.ref_name when no custom tag provided
- Enable manual release triggers with specified tag parameter