Repository navigation
Conversation
…pact view A look into saved output showed its raw handle while the model streamed the call in, and each search of one output took a row of its own that repeated the source in full. The provider's finished call now reaches the transcript with its arguments while it waits to start, so its row has its finished words from the start, and searches of one output share a row that names the source by its short name and each search once. MCP rows name the server and then the tool. A toolbox open of a skill counts as a skill in its group's header. A command the safety review held reads "Held <command> · safety caution", and a question fx stopped before it was answered reads "Couldn't ask you · fx stopped" after a resume. Ctrl+O, ACP, fx ask, and what the model gets stay as they were.
Another program's MCP tool name, such as mcp__filesystem__read_text_file, has no server before its first "_", so its row lost its name. Such a row now keeps the name core gives it.
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Searching command output for "FUZZ_", thenSearched command output for "FUZZ_";Opening tool call T5, thenOpened tool call T5. Ctrl+O still shows a call's arguments only once it starts.+Nrule, searches of different outputs keep a row each, and an output that no call in the transcript saved goes by core's noun (command output,saved output).Ran MCP mcp_slack_slack_read_thread. They now name the server, then the tool, without a tool name's repeat of its server's name:Ran slack read_thread,Running bloop get_account,Couldn't run fixture echo. The server comes from the call's result, which names it first. A call with no result yet takes the alias up to its first_.opened a skillin its group's header. A command the safety review held readsHeld python3 · safety cautioninstead ofSafety caution python3. A question fx stopped before it was answered reads● Couldn't ask you · fx stoppedafter a resume instead of● Failed.Only the compact view changes. Core's status text, Ctrl+O, ACP,
fx ask, and what the model gets stay as they were. Core keeps the words it shares with those surfaces, and every name the compact row shows is terminal-safe (R1.3, R4.6). Each fix has a unit test and an E2E test that fail onmain(R7.1).