feat(desktop): improve tool call detail views - #69868
Merged
Merged
Conversation
Contributor
૮ >ﻌ< ა ci reviewran on 06ee9d4 ℹ️ InfoDesktop E2E visual evidence · View test artifacts · View job1 visual diff. inline evidence is publishing... |
alt-glitch
added
type/feature
Show failed image-generation calls through the normal fallback row, remove duplicate normal-mode web-search JSON, and format Technical Mode payloads as readable JSON.
Show terminal commands with a prompt and exit status, then reveal ANSI-safe stdout and stderr in the expanded tool row.
Match context-only tool starts with command-bearing completions when their IDs differ, preventing stale duplicate terminal rows. Show the web-search query above its result cards.
ethernet8023
force-pushed
the
ethie/tool-call-nicer
branch
from
July 23, 2026 05:15
0d8f525 to
06ee9d4
Compare
Digidash23
added a commit
to Digidash23/hermes-agent
that referenced
this pull request
Jul 23, 2026
Ports upstream's feat(desktop): improve tool call detail views (NousResearch#69868, 26f1f6a) into apps/ccf, which has its own diverged copy of this rendering (our tool-diffs store shape differs from desktop's). Isolated the exact diff this one commit made (not the accumulated drift between the two apps) and applied it precisely: - fallback-model/types.ts + index.ts: new searchQuery/terminalCommand/ terminalExitCode view fields, shellCommand() helper extracted - fallback.tsx: terminal tool calls now render as a transcript ($ prompt + exit code badge, expandable to ANSI-safe stdout/stderr) instead of raw JSON; Technical Mode payloads pretty-print with JSON.stringify(…, null, 2) and labeled Arguments/Result sections instead of compact unlabeled JSON; web-search results show their query above the hit list - message-parts.tsx: a failed/malformed image-generation result now falls through to the normal tool row (readable error) instead of the image card silently rendering nothing - chat-messages.ts: tool-call reconciliation now also matches on command/code/path, not just search_term/query/question — a context-only tool-start and its command-bearing completion correlate correctly even when their IDs differ, preventing stale duplicate terminal rows Verified live: ran a real terminal command through CCF, confirmed the transcript renders with the $ prompt and exit code badge.
randlee
pushed a commit
to randlee/hermes-agent
that referenced
this pull request
Aug 11, 2026
* fix(desktop): improve fallback tool-call details Show failed image-generation calls through the normal fallback row, remove duplicate normal-mode web-search JSON, and format Technical Mode payloads as readable JSON. * feat(desktop): render terminal tool calls as transcripts Show terminal commands with a prompt and exit status, then reveal ANSI-safe stdout and stderr in the expanded tool row. * fix(desktop): reconcile tool calls by command Match context-only tool starts with command-bearing completions when their IDs differ, preventing stale duplicate terminal rows. Show the web-search query above its result cards.
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
2 participants
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.
What does this PR do?
Improves Hermes Desktop tool-call detail views so useful product UI does not get buried under raw payloads:
Related Issue
No related issue.
Type of Change
Changes Made
apps/desktop/src/components/assistant-ui/tool/fallback.tsx: simplify normal tool detail views, format Technical Mode payloads, and render terminal command/exit metadata.apps/desktop/src/components/assistant-ui/tool/fallback-model/{index,types}.ts: expose terminal command and exit-code presentation data.apps/desktop/src/components/assistant-ui/thread/message-parts.tsx: route malformed/failed image generation results to the fallback row.How to Test
nix develop -c npm run check --workspace=apps/desktop.$ command, exit code, and stdout/stderr render.Checklist
Code
fix(scope):,feat(scope):, etc.)pytest tests/ -qand all tests passDocumentation & Housekeeping
docs/, docstrings) — or N/A: no user-facing docs change is neededcli-config.yaml.exampleif I added/changed config keys — or N/ACONTRIBUTING.mdorAGENTS.mdif I changed architecture or workflows — or N/AScreenshots / Logs
Captured from a seeded Desktop session in
WLR_BACKENDS=headless cagevia Playwright. Each image is pinned to evidence commitd2b28403d5b75893548c5e96f1136e5a3b8fb60f.Terminal — one reconciled completed row, command transcript, stdout, and exit status
Web search — query at the top, above structured results
Image generation failure — readable fallback instead of a blank result
Technical Mode — indented payload arguments for every changed row
Verification
nix develop -c npm run check --workspace=apps/desktop: 2,649 passed, 3 skipped; renderer/Electron production builds and Linux unpacked package passed.nix develop -c npm run test --workspace=apps/desktop -- --run src/lib/chat-messages.test.ts src/components/assistant-ui/tool/fallback-model.test.ts: 63 passed.nix develop -c env WLR_BACKENDS=headless cage -- npx playwright test e2e/pr-tool-call-evidence.spec.ts.