Commit Graph

398 Commits

Author SHA1 Message Date
Jack Amadeo 2eca4b8416 MCP session replay integration test (#3939) 2025-08-12 14:03:49 -04:00
Amed Rodriguez b0d3c7ddad sanitize message content on deserialization (#3966) 2025-08-12 10:02:37 -07:00
Michael Neale 168366be3b fit: default ollama port (#4001) 2025-08-12 12:16:31 +10:00
Jack Amadeo dd504741a3 Remove cognitive complexity clippy lint (#4010) 2025-08-11 20:24:37 -04:00
Kat Hawthorne c2859fd49b feat: add @-mention file reference expansion to .goosehints (#3873)
Summary
Adds support for @filename references in .goosehints files, enabling dynamic inclusion of documentation files with security safeguards.

Changes
Add file reference parsing with regex pattern matching (@README.md, @docs/api.md, etc.)
Implement secure recursive file expansion with path traversal protection
Add ReDoS protection (128KB limit) and circular reference detection
Integrate with existing .gooseignore pattern system
Add comprehensive test suite covering security, functionality, and edge cases
Merge with latest upstream improvements (151 commits ahead)
2025-08-11 16:07:22 -07:00
tlongwell-block 566b9dc577 To-Do Tools (#3902)
Co-authored-by: David Katz <dkatz@squareup.com>
2025-08-11 15:57:17 -04:00
Rajiv Singh ab0b078314 feat: Add GOOSE_TERMINAL env variable to spawned terminals (#3911)
Signed-off-by: iamrajiv <rajivperfect007@gmail.com>
2025-08-11 16:08:45 +10:00
Amed Rodriguez 48c9af0ed9 Sanitize Tags Unicode Block (#3920) 2025-08-08 09:06:08 -07:00
Douwe Osinga f1b833f977 Move hardcoded LLM prompts to template files (#3934)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
Co-authored-by: Claude <noreply@anthropic.com>
2025-08-08 17:06:34 +02:00
Michael Neale e124bb084f feat: first time automated ollama install experience and openrouter (#3881)
Co-authored-by: Zane Staggs <zane@squareup.com>
2025-08-08 11:41:53 +10:00
Matt Nicholson 16304121ef add gpt-5 to openai provider format (#3924) 2025-08-07 17:19:39 -04:00
Zane e5e6522942 added gpt5 context limit (#3927) 2025-08-07 13:31:22 -07:00
David Katz 0c28228c1e Bump auto-compact threshold to 80% (#3925) 2025-08-07 15:26:20 -04:00
Angela Ning 6e6e0834e7 FIX: gemini tool call hanging (#3898) 2025-08-07 11:58:22 -04:00
Jack Amadeo 5eb01004a4 chore: use a Conversation type (#3735) 2025-08-06 22:06:13 -04:00
Lifei Zhou 67e63624ab fixed recipe schedule jobs (#3880) 2025-08-07 11:18:39 +10:00
Max Novich 4567eb22be additional metrics + Ui implementation (#3871) 2025-08-06 09:26:15 -07:00
Adam Tarantino c2f08557af feat: add OAuth provider abstraction for CLI configuration (#3157)
Signed-off-by: Adam Tarantino <tarantino.adam@gmail.com>
2025-08-06 17:04:49 +02:00
Jack Amadeo 6b93260fd0 Use RMCP for StreamableHTTP OAuth support (#3845) 2025-08-06 09:02:38 -04:00
Jack Amadeo ee450254b6 Try to keep key order for Databricks (#3876)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-08-06 09:02:10 -04:00
Rahul Zhade 7f877960d3 Fix OpenAI Provider with GitHub Models (#3875)
Signed-off-by: Rahul Zhade <rzhade3@users.noreply.github.com>
2025-08-06 11:45:08 +02:00
Jack Amadeo 7174fc74c3 Detect client disconnects and cancel tool calls (#3782) 2025-08-05 11:48:15 +02:00
David Katz 97e508b3a1 Fix leaky env variable causing flaky test (#3761) 2025-08-05 10:45:04 +02:00
Douwe Osinga 66f8ef123b Update gemini error msg (#3847)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
2025-08-05 10:40:21 +02:00
Douwe Osinga 918faddf03 Generic retry and error parsing (#3558)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-08-05 10:37:04 +02:00
David Katz 63f4374820 Token counting in Auto-compact uses provider metadata (#3788) 2025-08-04 15:45:23 -04:00
Jack Amadeo f7de3596cf fix param order of debug_conversation_fixer (#3796) 2025-08-04 12:13:50 -04:00
Colin Walters 412ceb8128 chore(mcp): Add helpers to parse parameters (#2821)
Signed-off-by: Colin Walters <walters@verbum.org>
2025-08-03 10:51:36 +02:00
Angela Ning d93063b686 Override session name generator for ollama provider (#3710) 2025-08-02 19:36:21 +02:00
Jack Amadeo 6b3018b77a Swap MCP client implementations to rmcp (#3634) 2025-08-01 17:14:41 -04:00
Max Novich 3b6a92b715 Mnovich/otlp v2 (#3772) 2025-08-01 12:15:12 -07:00
Douwe Osinga 750fbb44fb No tool role means we should not collapse messages (#3778)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
2025-08-01 16:53:46 +02:00
Andrew Harvard 900698777b integrate MCP UI (#2948)
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-08-01 08:09:06 -04:00
Jack Amadeo b4ab774329 Fix claude model names (#3765)
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-08-01 13:33:08 +10:00
Michael Neale bbf1f1eee8 fix: don't return full shell output when very large (#3750)
has some additional build fixes
2025-08-01 12:48:25 +10:00
Shodipo Ayomide 26cb85094f Increase req body limit (#2965)
force merging this since it has been going back and forth for so long
2025-07-31 22:23:09 +02:00
David Katz 9daa464021 Env parsing for primitive types (#3706) 2025-07-31 14:51:03 -04:00
David Katz 6f36c1f81e Autocompact + One Shot Summarization algorithm (#3559)
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-07-31 14:23:06 -04:00
Lifei Zhou e20a88516d chord: refactor tool route (#3732) 2025-07-31 08:39:00 +10:00
Douwe Osinga 21604df685 Ok, well, that got out of hand (#3718)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
Co-authored-by: Michael Neale <michael.neale@gmail.com>
2025-07-30 11:41:42 +02:00
Michael Neale 601c678722 feat: openrouter out of the box experience for goose installations (#3507) 2025-07-30 17:17:43 +10:00
Lifei Zhou 1544e1fb27 chore: small refactor on agent.rs (#3703) 2025-07-30 09:22:19 +10:00
Douwe Osinga eb0048018f Fix tool vector tests (#3709)
Co-authored-by: Douwe Osinga <douwe@squareup.com>
2025-07-29 19:31:29 +02:00
Kenneth Mungai 9d996c3957 feat: update Groq models for better tool calling support (#3676)
Co-authored-by: kenneth.kimani <kenneth.kimani@savannahinformatics.com>
Co-authored-by: Douwe Osinga <douwe@block.xyz>
2025-07-29 16:33:21 +02:00
Julian Brown bc25308d7b fix: escape control characters in LLM tool call arguments JSON (#2893)
Signed-off-by: Julian Brown <contact@julianbrown.dev>
Co-authored-by: Julian Brown <jbrown@stripe.com>
2025-07-29 11:05:01 +02:00
Gustavo Hoirisch c84dc392be feat(githubcopilot): add ability to fetch supported models (#2717)
Signed-off-by: Gustavo Hoirisch <gugahoi@users.noreply.github.com>
2025-07-29 11:01:18 +02:00
Jack Amadeo 71d5ae4961 Create a message ID for tool response messages (#3591) 2025-07-29 10:25:12 +02:00
Jarrod Sibbison 5de9f84619 fix: Ensures final output tool is available when using vector tool search (#3701) 2025-07-29 12:18:46 +10:00
Michael Neale 18be07853b chore: adding in some new models token limits (#3685) 2025-07-29 12:12:21 +10:00
Michael Neale 2382763568 fix: have option for cli providers to use their configured or default model (#3683) 2025-07-29 10:27:53 +10:00