Skip to content

Rebuild MCP search and receipt tools#1375

Merged
r3dbars merged 7 commits into
mainfrom
codex/mcp-search-rebuild-20260703
Jul 3, 2026
Merged

Rebuild MCP search and receipt tools#1375
r3dbars merged 7 commits into
mainfrom
codex/mcp-search-rebuild-20260703

Conversation

@r3dbars

@r3dbars r3dbars commented Jul 3, 2026

Copy link
Copy Markdown
Owner

Summary

Checks

  • bash build-deps.sh --force
  • bash build.sh --no-open
  • bash run-tests.sh — 11,758 passed
  • bash run-integration-smoke.sh
  • swift test — 590 passed, 2 skipped
  • swift test --package-path Tools/TranscriptedMCP — 157 passed
  • git diff --check origin/main...HEAD
  • after merging latest main: swift test --package-path Tools/TranscriptedMCP — 157 passed; git diff --check origin/main...HEAD

Stale PRs

COORD_DONE: GREEN | PRs opened: this draft replacement | stale PRs salvaged/held: #1368/#1369/#1365 salvaged here; #1363/#1373 already merged to main | checks run: build-deps, build, tests, integration smoke, swift test, MCP swift test, diff-check | lanes used: Codex=rebuilt/resolved/tested/opened PR; Claude=skipped unavailable (not logged in); Local=attempted triage but output unusable, proof /Users/redbars/.codex/maestro/runs/20260703T105917Z-transcripted-mcp-pr-triage-local; Windows=skipped unavailable | smallest next action: review draft PR and merge only after CI/human review

claude and others added 7 commits July 3, 2026 06:04
…read_dictation

The two read tools were token bombs: read_meeting returned an entire
90-minute transcript verbatim and read_dictation dumped the whole day
file. Both now support optional offset/limit params and a ~30k-char size
guard (maxUnpaginatedReadCharacters): anything paginated or over budget
comes back as a JSON window with total_utterances/total_entries, offset,
returned, truncated, next_offset, and a continuation hint. Frontmatter
metadata is preserved on full-section meeting pages; utterance windows
reuse the TranscriptedCaptureKit parser. Small unpaginated reads stay
byte-identical raw markdown, and entry_id reads are unchanged.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GuGZaFRmNrqfGPpqf7WH4n
Three test fixtures built labeled tuples with inline Double math inside
map closures; the Swift type checker times out on that expression shape
(build-and-test failure on CI). Replace them with a shared loop-based
makeSequentialUtterances helper in TestHelpers.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GuGZaFRmNrqfGPpqf7WH4n
…ild-20260703

# Conflicts:
#	Tools/TranscriptedMCP/CLAUDE.md
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants