chore(release): 0.14.0

Release 0.13.0

This release adds new session commands and improved task management features, enhancing user interaction and planning capabilities (/run, /prompt, /plan) (61f22cd, e32c88b, 1be66ad). Multimodal support and context handling are improved for smoother workflows (c33d653, 8b3876d, 36263e3). Several bug fixes address Windows path handling, API cost tracking, and tool response errors, improving overall stability and usability (9a9c656, 71aa15f, dd41044). Documentation and installation guides have also been updated for better user guidance (d842678, 9814b74).

- **session**: add completion and hints for /run and /prompt `61f22cda`
- **utils**: add context block detection and expansion `c33d653e`
- **chat**: add Ctrl+G to add message without sending `8b3876de`
- **mcp**: require tasks with title and description `e32c88be`
- **plan**: add /plan command to display current plan status `1be66add`
- **truncation**: add global MCP response tokens threshold `36263e31`

- **file_parser**: add utilities for file reference extraction and rendering `041b8c46`

- **utils**: preserve Windows drive letters in file path parsing and rendering `9a9c656e`
- **utils**: normalize file paths on Windows in read_file_lines `92543735`
- **prompt**: prevent accidental continuation trigger on /prompt command `848f8293`
- **session**: preserve system message on /run command replace output `d78631c0`
- **session**: track API call cost immediately after response `71aa15f1`
- **cargo**: upgrade dependencies to fix cargo audit issues `6a144ece`
- **fs**: suggest line_replace for ambiguous replacements `3904a5a7`
- **mcp**: handle error flag in MCP tool responses correctly `dd410443`

- **config**: update configuration docs with command groups `d8426784`
- **utils**: add file parsing and rendering usage instructions `228c86bc`
- **installation**: rewrite installation guide `9814b74e`
- **plan**: enhance task description guidelines with examples `23b7934b`

- **file_renderer**: fix Windows test failures by normalizing paths and line endings `e57f5aba`
- **plan**: fix task format in plan tests after refactor `b829ef4f`

**Total commits**: 21 across 5 categories

✨ **6** new features - *Enhanced functionality*
🔧 **1** improvement - *Better performance & code quality*
🐛 **8** bug fixes - *Improved stability*
📚 **4** documentation updates - *Better developer experience*
🔄 **2** other changes - *Maintenance & tooling*

Release 0.12.0

This release adds reusable prompt templates and spending tracking to enhance session management, along with batch processing for improved tool efficiency (9848036, f5b4620, 5a3926f). Several bug fixes improve search accuracy, prevent retry loops, and enhance system stability (a24932e, b9d4f78, 4edc057). Dependency updates and testing refinements further optimize overall performance (a69f72b, 74d8051).

- **mcp,chat**: batch large output prompts for parallel tool calls `5a3926f3`
- **session**: add /prompt command with reusable templates `9848036f`
- **session**: add request spending threshold and tracking `f5b46206`

- **web**: clarify no-results issue with multiple quoted phrases `a24932e7`
- **style**: update format strings for Clippy compliance `01aaacf0`
- **list_files**: treat content pattern as fixed string to avoid regex errors `efb8a3b5`
- **session**: prevent infinite retry loop on continuation failure `b9d4f783`
- **mcp**: prevent recursion in cancellation polling loop `4edc057f`

- **deps**: update and reorganize dependencies `a69f72b9`
- **fs**: reset line count tracking in batch_edit test `74d80515`

**Total commits**: 10 across 3 categories

✨ **3** new features - *Enhanced functionality*
🐛 **5** bug fixes - *Improved stability*
🔄 **2** other changes - *Maintenance & tooling*

Release 0.11.0

This release adds support for the latest GPT-5 and Anthropic opus-4-1 AI models, expanding the range of AI capabilities (076e797, b8f3576). Several improvements enhance session stability and usability, including better input rendering with ANSI color support and cancellation options for agent tools (574d08e, ed9e282, e245ecf, a07109d). Additionally, bug fixes prevent infinite retry loops and improve text editing reliability (574d08e, 315ddfd).

- **openai**: add GPT-5 model support with pricing and params `076e797d`
- **anthropic**: add opus-4-1 model and fix temp/top_p handling `b8f3576b`

- **session**: prevent infinite retries on continuation calls `574d08e5`
- **text_editing**: prevent repeated line_replace on new lines `315ddfda`
- **mcp**: add cancellation support to agent tool execution `ed9e2820`
- **session**: enable ANSI color mode for rustyline input rendering `e245ecff`
- **session**: prevent infinite retry loops on continuation errors `a07109dd`

**Total commits**: 7 across 2 categories

✨ **2** new features - *Enhanced functionality*
🐛 **5** bug fixes - *Improved stability*

Release 0.10.2

This release improves session reliability by addressing incomplete tool calls, message truncation, and session resuming issues (105fcd8, 2ba6773, a281b8d, 54aed7c, 0ace265, d9ac1b3, 7b5dfad). Additionally, the /done command has been streamlined for better usability, and new quick start and troubleshooting guides have been added to enhance user onboarding (855d189, 9872b47).

- **done**: move /done command to dedicated file and clean code `855d1898`

- **session**: detect and truncate earliest incomplete tool calls on r... `105fcd80`
- **session**: truncate messages on interrupted tool calls to clean state `2ba67733`
- **session**: correct tool_calls reconstruction on session resume `a281b8d5`
- **session**: restore layers state and cost on session resume `54aed7c5`
- **session**: handle incomplete tool calls in session resuming `0ace2654`
- **session**: re-add initial messages on /done command completion `d9ac1b30`
- **session**: prevent infinite loop on Ctrl+C cancellation `7b5dfada`

- **instructions**: add quick start and troubleshooting guide `9872b47f`

**Total commits**: 9 across 3 categories

🔧 **1** improvement - *Better performance & code quality*
🐛 **7** bug fixes - *Improved stability*
📚 **1** documentation update - *Better developer experience*

Release 0.10.1

This release addresses several session management issues to improve stability and user experience, including better handling of cancellation signals and cleanup of partial messages (61a235d, 959310f, 7e5ec3e). Documentation has also been updated to reflect these enhancements.

- **session**: resolve cancellation issue and update documentation `61a235d4`
- **session**: handle cancellation signal correctly in tool loop `959310fb`
- **session**: clean up partial messages on tool and layer cancellation `7e5ec3ea`

**Total commits**: 3 across 1 categories

🐛 **3** bug fixes - *Improved stability*

Release 0.10.0

This release adds customizable output roles for session messages to enhance interaction control (ca1e288). Several bug fixes improve session stability, including better cancellation handling, prevention of duplicate messages, and stricter configuration enforcement, alongside file lock implementation to avoid write conflicts and improved tool validation (93821f6, 6a9f835, d25a7cd, e2a05cc, d2e707f, e727776, 5379398). Additionally, the session flow has been streamlined for a smoother user experience (4091cf2).

- **session**: add output_role control for session messages `ca1e288f`

- **session**: simplify interactive and non-interactive session flow `4091cf29`

- **session**: propagate Ctrl+C cancellation to animation and tools `93821f6c`
- **mcp/fs**: add file locks to prevent concurrent write conflicts `6a9f835e`
- **config**: add missing output_role to default config sections `d25a7cd7`
- **session**: prevent duplicate user messages with layers and continu... `e2a05ccc`
- **session**: remove defaults to enforce strict output role config `d2e707f9`
- **session**: replace ctrlc crate with tokio signal for Ctrl+C handling `e7277764`
- **layers**: apply pattern-based tool validation in layers `5379398d`

**Total commits**: 9 across 3 categories

✨ **1** new feature - *Enhanced functionality*
🔧 **1** improvement - *Better performance & code quality*
🐛 **7** bug fixes - *Improved stability*

Release 0.9.0

This release introduces enhanced session management with dynamic role switching and customizable prompts and temperature settings for improved AI interactions (38ee56f, 0bdaed2, 5be4bbc). Several bug fixes improve configuration stability and reliability, while additional tests ensure robust handling of batch edits (3111ac2, 20301e7).

- **config**: add top_k and top_p defaults and tune temperatures `38ee56f2`
- **config,ask,shell**: add configurable prompts and temperatures for... `0bdaed2b`
- **session**: add /role command to switch session role at runtime `5be4bbc5`

- **config**: use slice instead of Vec reference in show_mcp_servers `3111ac2f`

- **fs**: add critical batch_edit tests for line number handling `20301e74`

**Total commits**: 5 across 3 categories

✨ **3** new features - *Enhanced functionality*
🐛 **1** bug fix - *Improved stability*
🔄 **1** other change - *Maintenance & tooling*

Release 0.8.1

This release enhances session management with improved task finalization, streamlined continuation handling, and more reliable command processing (5523f0d, cf96448, 6ac8362, bb2f57a). User experience is improved through clearer guidance on parallel tool usage and more intuitive command completion (b328fda, 02e2025). Several bug fixes address token handling, logging clarity, and error reporting to ensure smoother and more stable interactions (5fc894a, fac0836, 2e24515, 58da6f3, ef4eb89).

- **session**: simplify continuation logic and unify triggers `feb822c0`
- **session**: remove unused syntax highlighter methods and tests `93727213`

- **session**: update /done command message to "Finalizing current task `5523f0da`
- **session**: remove duplicated token limit log message `5fc894ac`
- **session**: use root max_retries instead of role config `fac08362`
- **session**: respect CLI options over config defaults for runtime pa... `2e245158`
- **session**: simplify continuation logic and avoid user prompt on to... `cf964480`
- **session**: handle continuation immediately to fix tool bug `6ac8362e`
- **session**: stop tool processing on continuation trigger `bb2f57a0`
- **session**: return error on invalid session token threshold at launch `58da6f35`
- **session**: run continuation check before tool calls evaluation `ef4eb893`
- **session**: enable bash-like completion with partial matches and re... `02e20255`

- **session**: clarify guidance on parallel tool usage in continuation `b328fdad`

**Total commits**: 13 across 3 categories

🔧 **2** improvements - *Better performance & code quality*
🐛 **10** bug fixes - *Improved stability*
📚 **1** documentation update - *Better developer experience*

Release 0.8.0

This release introduces enhanced session management with improved token tracking, continuation controls, and time monitoring, alongside expanded developer tools and multimodal support (6fabe56, b0ce465, 3b81c7f). Configurable retry logic and updated pricing models improve AI provider integrations, while comprehensive documentation and testing bolster usability and reliability. Several bug fixes address token calculation accuracy, error handling, and session stability to ensure a smoother user experience (41090fc, c79cfb5, f7270ae).

- **session**: add role-based token counting and continuation checks `6fabe563`
- **session**: include system prompt and tools in context token count `b0ce4657`
- **session**: add flag to disable continuation triggers temporarily `174ca557`
- **deepseek**: update pricing scheme to use hash maps and helpers `d6ceb6e8`
- **dev**: add plan tool to developer MCP server `3b81c7fd`
- **session**: add API, tool, and total time tracking in layers `d5fcb59b`
- **fs**: support negative line ranges in text editor view_range `63778fed`
- **api**: add configurable retry logic for Amazon provider `2885fb13`
- **session**: preserve initial instructions and welcome messages on ... `880ed2d8`
- **batch_edit**: support single-file multiple operations with origin... `b497a2ca`

- **providers**: unify Anthropic and OpenAI retry logic using ret... `ad111d0f`
- **agent,session**: source temperature from role config instead ... `d792da03`
- **batch_edit**: extract batch_edit as independent tool from tex... `2ac3c987`
- **ci**: fix markdown code block formatting in release workflow `0f2fde3e`

- **anthropic**: correct token usage calculation including cache tokens `41090fcd`
- **config**: track and display env var source including .env override `8f3805db`
- **session**: resolve OpenAI 400 errors and add CTRL-C cancellation `c79cfb5f`
- **plan**: prevent overwriting active plan on start command `4bd44337`
- **session**: correct continuation trigger timing in response processing `f513f964`
- **openai**: correct token cost calculation with cache tokens `10716178`
- **openai**: extract and set tool_call_id from response `1d76a461`
- **chat**: use correct model and params for auto threshold continuation `f7270aee`

- **mcp**: add detailed docs for new plan tool and usage `f597b7c0`
- **doc**: actualize installation, overview, and config guides `8dc57fce`
- specify cargo check with short message format in instructions `24862099`

- **plan**: fix async test assertions and add serial execution `d44b768a`
- **plan**: add comprehensive tests for plan tool commands `3f49e94a`
- **anthropic**: update model pricing to June 2025 rates `4f558d96`
- **openai**: update model list and pricing to 2025 versions `e9f0841f`

**Total commits**: 29 across 5 categories

✨ **10** new features - *Enhanced functionality*
🔧 **4** improvements - *Better performance & code quality*
🐛 **8** bug fixes - *Improved stability*
📚 **3** documentation updates - *Better developer experience*
🔄 **4** other changes - *Maintenance & tooling*