Extend the sensitive-key decoration to cover the full logical value, not
just the rest of the current line. Adds value-extent detection for:
- TOML triple-quoted strings (multi-line literal/basic strings)
- TOML arrays ([ ... ] across multiple lines, string-aware)
- YAML block scalars (| and > with indent-based termination)
- JSON arrays and objects (balanced braces, string-aware)
Without this, a secret value spanning multiple lines (multi-line API
key, certificate, private key) was only blurred on its opening line and
the rest stayed readable when the Eye toggle was masked.
Adds regression tests for the TOML triple-quote and array cases.
Flagged by PR-Agent review on #1248.
Total dist/ui gzipped grew from ~1.5MB to ~1.6MB after the CodeEditor
switch from react-simple-code-editor + prism-react-renderer to
CodeMirror 6. The verify-bundle script preamble notes the cap is a
sanity check for accidental large dependencies, not a hard
performance target; this is an intentional editor upgrade to restore
native selection/copy in the codex config.toml viewer.
Replace the hand-rolled textarea+Prism overlay in CodeEditor with
CodeMirror 6. The overlay hid the textarea selection with
`selection:bg-transparent` and `WebkitTextFillColor: transparent` as a
workaround for wrap drift between the highlight layer and the textarea,
which left Ctrl+A / drag-select / Ctrl+C with no visible result.
CodeMirror handles selection, copy, undo, and wrap natively. Custom
themes match the previous prism-react-renderer github (light) and
nightOwl (dark) palettes; background is transparent so the surrounding
bg-muted/30 still shows; font, font-size, line-height, padding, and
gutter-less layout are preserved.
Public CodeEditor props are unchanged (value, onChange, language,
readonly, exactText, className, minHeight, heightMode) so all seven
call sites keep working. The sensitive-key blur is reimplemented as a
CodeMirror Decoration and stays gated behind the existing Eye toggle.
Test setup gets a Range.getClientRects polyfill so CM mounts cleanly in
JSDOM, and shared-page.test.tsx checks .cm-content text directly.
Closes#1247
Squash merge PR #1240.\n\nValidated locally from the PR head because trusted-author GitHub CI was skipped for this fork PR:\n- bun run validate\n- bun run build:all\n- bun run test:all\n- CCS_E2E_SKIP_BUILD=1 bun run test:e2e
Closes#1218.
When the user invokes `ccs <profile> agents` (and other Claude subcommands
like `mcp`, `doctor`, `plugin`, ...), CCS was unconditionally injecting
session-only Claude flags (`--append-system-prompt`, `--disallowedTools`,
`--settings`, official-channels plugin specs) and forwarding the
`DISABLE_TELEMETRY=1` env var from profile settings. Each of those
either errored out the subcommand (`error: unknown option
'--append-system-prompt'`) or silently flipped Claude into
non-interactive list mode, so the new `claude agents` agent view never
opened under CCS.
Add a small `claude-subcommand-detector` that recognizes the documented
Claude subcommand set after skipping known value-taking flags, then have
the three steering-prompt injectors (websearch, image-analysis, browser),
the cliproxy and settings launchers, and the official-channels plan
short-circuit when a subcommand invocation is detected. Also strip
`DISABLE_TELEMETRY` from the spawned env only for subcommand
invocations — upstream Claude Code uses that var as a kill switch for
the subcommand TUIs, and the user's telemetry preference still applies
to every normal interactive session.
Verified end-to-end against `ccs glm agents` and `ccs ck agents`: the
agent view opens correctly, no `unknown option` error, and all 1938
existing tests pass.
Keep active CCS workflows on self-hosted runners, gate self-hosted PR execution to trusted authors, and scope privileged release credentials to the exact git operations that need them.
Tam Nhu Tran