feat: v2.4 Phase 30 — Claude Code CLI Harness#18
Merged
RichardHightower merged 31 commits intomainfrom Feb 26, 2026
Merged
Conversation
- 9 valid fixtures: SessionStart, UserPromptSubmit, PreToolUse, PostToolUse, AssistantResponse, SubagentStart, SubagentStop, Stop, SessionEnd - 1 malformed fixture for negative testing - All fixtures match CchEvent struct in memory-ingest Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- 5-CLI matrix (claude-code, gemini, opencode, copilot, codex) x 2-OS - Bats-core installation with bats-support and bats-assert libraries - JUnit XML report generation via --report-formatter junit - Failure artifact upload with 7-day retention - Skip annotation when CLI test directory not found Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Workspace isolation with setup_workspace/teardown_workspace - Daemon lifecycle: build_daemon_if_needed, start_daemon, stop_daemon - Health checking with grpcurl, nc, and /dev/tcp fallbacks - Random port selection for test isolation - grpc_query and ingest_event helpers - Auto-detect PROJECT_ROOT via workspace Cargo.toml Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- CLI detection: require_cli (skip) and has_cli (boolean) - Claude Code wrappers: run_claude, run_claude_with_hooks with timeout - Hook pipeline testing: run_hook_stdin, run_hook_stdin_dry - Cross-platform timeout detection (timeout/gtimeout) - .gitignore excludes .runs/ workspace artifacts Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- SUMMARY.md with self-check passed - STATE.md updated: plan 02 complete, progress 50% Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ingest - 6 always-run tests: daemon binary, ingest binary, health check, valid/malformed/empty JSON ingest - 2 claude-dependent tests: binary detection and headless JSON output (skip if absent) - Fix IPv6->IPv4 mismatch: daemon binds 0.0.0.0 but helpers used [::1], switched to 127.0.0.1 - Fix health check: daemon lacks grpc.health service, prefer nc TCP check over grpcurl - Add nested-session skip for CLAUDECODE env var Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…fication - 9 individual event type tests: SessionStart, UserPromptSubmit, PreToolUse, PostToolUse, AssistantResponse, SubagentStart, SubagentStop, Stop, SessionEnd - 1 multi-event sequence coherence test (4 events, same session) - Two-layer proof: Layer 1 (ingest exit 0 + continue:true), Layer 2 (gRPC query) - Unique session IDs per test via PID to avoid cross-test interference - Build-resilience: fallback to existing binary when cargo build fails - Fixture JSON rewritten with jq (sed fallback) for session_id isolation Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…tion - 5 tests covering complete session lifecycle, TOC browse, cwd metadata, real Claude Code hook fire, and concurrent session isolation - Fix: DEFAULT_ENDPOINT changed from [::1] to 127.0.0.1 to match daemon bind addr - Fix: Remove short flag from global --log-level to avoid clap conflict with --limit Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- 7 tests: daemon down, malformed JSON, empty stdin, unknown event type, timeout enforcement, wrong port, and large payload handling - All tests verify fail-open behavior (exit 0, continue:true) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- SUMMARY.md with 2 task commits, 3 deviation fixes documented - STATE.md updated: phase 30 plan 4/4 complete, progress 68% Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Read MEMORY_DAEMON_ADDR from environment before gRPC connect - Use MemoryClient::connect(addr) when env var is set - Fall back to connect_default() (port 50051) when unset - Preserves fail-open behavior throughout Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Remove PIPELINE_PORT=50051 variable - Call start_daemon with no args (uses random port via pick_random_port) - Remove outdated NOTE about hardcoded DEFAULT_ENDPOINT - Update header comment to reflect OS-assigned random port isolation Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- SUMMARY.md documents env var support and random port isolation - STATE.md updated with plan position, decisions, metrics Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
… in hooks.bats - All 10 hook capture tests now use hard Layer 2 assertions - Tests fail properly when session_id not found in gRPC query results - Content assertions added for tests 2-5 (project structure, Read tool) - Removed if-guard around result check -- empty result now fails as expected Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ase 30 - Mark plans 30-05 and 30-06 as complete in plan list - Update Phase 30 progress from 0/4 to 6/6 in progress table Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…tions Root cause: memory-ingest reads stdin line-by-line, so multi-line JSON from jq pretty-print silently failed (fail-open returns continue:true). Also, query output format never includes session_id. Changes: - Use jq -c (compact) in rewrite_session_id to ensure single-line JSON - Replace session_id assertions with event count and content assertions - Use wide time window (from=0) to avoid timestamp precision issues - All 30/30 bats tests now pass (hooks 10, smoke 8, pipeline 5, negative 7) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
RichardHightower
temporarily deployed
to
e2e-cli
GitHub Actions
Inactive
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
common.bash,cli_wrappers.bash) with workspace isolation, daemon lifecycle, random port selectione2e-cli.yml) with 5-CLI x 2-OS matrix (claude-code, gemini, opencode, copilot, codex)MEMORY_DAEMON_ADDRenv var support tomemory-ingestfor test port isolation-lfor--log-levelvs--limit)Key fixes in final commits
memory-ingestreads stdin line-by-line, so multi-line JSON fromjqpretty-print silently failed. Fixed withjq -c(compact output)[timestamp] agent_type: content— session_id is never shown. Replaced with count/content assertionsTest plan
bats tests/cli/claude-code/— 30/30 pass (1 expected skip for nested Claude)cargo clippy --workspace— zero warningscargo test --workspace— all pass🤖 Generated with Claude Code