Automated summary
Verifies integration coherence in quant.cpp by cross-comparing module boundaries (types ↔ blocks, quantize→attention pipeline, cache, progressive tiers, SIMD vs generic).
Strengths
- Provides five boundary checklists with concrete assertions (type sizes, ref_count, MSE, bit-exact equivalence)
- Scopes the review to the actual change set using git diff --name-only to identify affected modules
- Includes a historical bug-pattern table (NULL function pointers, hardcoded assumptions, ignored parameters, integer overflow) as a checklist
- Ends with an actionable build-and-test step (cmake --build followed by ctest)
Limitations
- Ships no test code itself: coverage depends on the repository's existing ctest suite
- Tightly coupled to quant.cpp's module layout and not reusable for other projects
- Thresholds and tolerances (MSE, attention similarity) are referenced without concrete numeric values
Use it before merging a change that touches quant.cpp's quantization types, cache, progressive compression, or SIMD kernels.
Avoid it on projects lacking this quant.cpp module structure, or where no build and test suite are available.
Security analysis
SafeThe skill is a read/verify-oriented QA checklist for a C++ quantization codebase. It contains no exfiltration, no destructive commands, no obfuscated payloads, no credential access, and no attempts to disable safety controls. The only executable action is a standard out-of-tree cmake build and ctest run, which is legitimate for integration verification, though it would run untrusted project build scripts — hence a minor note rather than a risk rating.
- •Instructions include running build/test commands (`cmake --build build && ctest --test-dir build`), which executes arbitrary project build scripts; this is normal for a QA skill but is a powerful action.
- •No allowed-tools declared, so the skill implicitly depends on whatever default tool set the agent has (bash/file read) rather than scoping it explicitly.
Examples
Value cache 쪽 코드를 수정했어. 머지 전에 git diff로 변경된 모듈 식별하고, cache→block 경계면 체크리스트랑 빌드/ctest까지 돌려서 정합성 검증해줘.quant.cpp 전체에 대해 QA 실행해줘: 타입 시스템↔블록, quantize→attention, cache→block, progressive→tier, SIMD→generic 다섯 경계면을 교차 비교하고 발견 사항 보고해줘.새 quantization 타입을 추가했는데 TQ_TRAITS 등록이 완전한지 검증해줘. attention 함수 포인터 NULL 여부와 type_size == sizeof(block) 정합성 확인 포함해서.name: qa description: "quant.cpp의 통합 정합성을 검증한다. 모듈 간 경계면 불일치, 타입 시스템 정합성, quantize→attention 파이프라인, cache 무결성을 교차 비교한다. '검증', 'QA', '테스트', '정합성 확인', 'validate' 요청 시, 또는 코드 변경 후 머지 전에 자동으로 사용."
QA — Integration Coherence Verification
"존재 확인"이 아닌 **"경계면 교차 비교"**가 핵심이다.
검증 대상: 5대 경계면
1. 타입 시스템 ↔ 블록 구조
TQ_TRAITS[type].type_size == sizeof(block_type)
TQ_TRAITS[type].attention != NULL (모든 7개 타입)
tq_type_bpe(type) == sizeof(block) * 8.0 / block_size
2. Quantize → Attention 파이프라인
for each type:
quantize(key) → attention(query, quantized) → finite score
quantize(key) → dequantize(quantized) → MSE < threshold
3. Cache → Block
cache_append(key, value) → cache_get_block() → valid pointer
cache_append(key, value) → cache_get_value() → valid pointer
cache_share_block() → ref_count == 2
4. Progressive → Tier
append N tokens → tier(0) for recent, tier(1) for warm, tier(2) for cold
recompression: warm_type → cold_type 정확한 타입 사용 확인
5. SIMD → Generic
neon_quantize(input) == generic_quantize(input) (bit-exact)
neon_attention(q, k) ≈ generic_attention(q, k) (tolerance)
실행 방법
- 변경된 파일 목록 확인 (
git diff --name-only) - 해당 파일이 속한 모듈 식별
- 해당 모듈의 경계면 체크리스트 실행
- 빌드 + 테스트:
cmake --build build && ctest --test-dir build - 발견 사항 보고
버그 패턴 (과거 발견 사례)
| 패턴 | 사례 | 교훈 | |------|------|------| | 함수 포인터 NULL | Uniform attention 미등록 (BUG-1) | traits 테이블 완전성 항상 검증 | | 하드코딩 가정 | Progressive가 UNIFORM_4B 가정 (BUG-2) | 동적 타입 조회 사용 | | 파라미터 무시 | Value cache 미저장 (BUG-3) | 모든 파라미터 사용 여부 확인 | | 정수 오버플로 | size 계산 (BUG-4) | 곱셈 전 오버플로 체크 |
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/integration-coherence-verification-mrwak8gk' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/integration-coherence-verification-mrwak8gk/download' --output '.skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md.part' && mv '.skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md.part' '.skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.
Claude Code
test -f '.skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md' && test -s '.skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/qa' && cp -R '.skill-downloads/integration-coherence-verification-mrwak8gk' '.claude/skills/qa'OpenSkills
OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.
Official OpenSkills installationopenskills install './.skill-downloads/integration-coherence-verification-mrwak8gk'openskills read 'integration-coherence-verification-mrwak8gk'Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.
For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.
openskills syncManual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/integration-coherence-verification-mrwak8gk/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
ClickHouse Performance Comparison
Testing
Evaluate ClickHouse performance test results, check PR performance changes, compare with history, and assess whether a result is actionable.
Analyze CI Performance Reports
Testing
Analyze CI performance comparison reports for a ClickHouse PR. Cross-references master history to distinguish real regressions from flaky tests.
Santa Method
Testing
Multi-agent adversarial verification with two independent reviewers. Output is checked twice and iteratively fixed until both pass before shipping.