Repository navigation
feat: optionally show reasoning and tool call details - #15
Merged
Merged
Conversation
Two Settings toggles, off by default, show the agent's reasoning and each tool call's input, output and error as collapsible plain text. Credential- like values are redacted and every field is clipped before it reaches chat state. Refs #13 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Recognise prefixed and camelCase secret names (GITHUB_TOKEN, client_secret, accessToken, x-api-key), JSON result text, Basic and Bearer values and user:pass in URLs, without hiding max_tokens or prose like "token expired". Reasoning text goes through the same redaction, nesting is capped, a completion without a name keeps the name from the start event, and the Settings hint now says redaction is best effort. Refs #13 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A pair such as `https:` or `Error:` no longer consumes the text after it, so `?token=`, `?api_key=` and `Error: NAME=VALUE` are redacted. Also recognise PGPASSWORD and MYSQL_PWD style names and lowercase `token` headers, while plain prose like "Basic authentication required" stays. Refs #13 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
thinking.deltatext). With tool details on, each tool chip opens into its input, output and error.<pre>; nothing is stored outside in-memory chat state.src/chat/tool-detail.ts): secret names in any case style (GITHUB_TOKEN,client_secret,accessToken,x-api-key,PGPASSWORD), secret CLI flags,?token=in URLs, JSON results field by field,Bearer/Basicvalues,user:pass@and well-known key prefixes. Each field is clipped to 4,000 characters. The hint and the threat model say redaction is best effort.Review
Isolated review ran twice. Both blocking findings (missed secret formats, harmless names swallowing later secrets) are fixed and covered by tests. Remaining non-blocking notes: details are not in server history, so a reload drops them; long reasoning is clipped at 4,000 characters.
Test plan
pnpm lint && pnpm typecheck && pnpm test(451 tests, with the ActionCard fixture fix from test: keep the command preview card open regardless of the date #10 applied locally)pnpm buildBase is
feat/8-ui-polish(#12) so the diff shows only this change.Closes #13
🤖 Generated with Claude Code