SQL Optimization
github/awesome-copilot
Universal SQL performance optimization assistant for comprehensive query tuning, indexing strategies, and database performance analysis across all SQL databases (MySQL, PostgreSQL, SQL Server…
SPIRV-Tools optimizer pass development, IR manipulation, and PassTest registration.
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install LuisaGroup/LuisaCompute spv-opt --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/spv-opt .claude/skills/spv-opt && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .claude/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-optType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install LuisaGroup/LuisaCompute spv-opt --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/spv-opt .agents/skills/spv-opt && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .agents/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install LuisaGroup/LuisaCompute spv-opt --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/spv-opt .cursor/skills/spv-opt && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .cursor/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/LuisaGroup/LuisaCompute.git --path .agents/skills/spv-opt--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install LuisaGroup/LuisaCompute spv-opt --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/spv-opt .gemini/skills/spv-opt && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .gemini/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install LuisaGroup/LuisaCompute spv-optInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/spv-opt .github/skills/spv-opt && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .github/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install LuisaGroup/LuisaCompute spv-opt --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LuisaGroup/LuisaCompute.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/spv-opt .opencode/skills/spv-opt && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "spv-opt" agent skill from https://github.com/LuisaGroup/LuisaCompute/tree/stable/.agents/skills/spv-opt into .opencode/skills/spv-opt/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "spv-opt", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
spv-optSPIRV-Tools optimizer pass development, IR manipulation, and PassTest registration.
Spv Opt is an agent skill from LuisaGroup/LuisaCompute. SPIRV-Tools optimizer pass development, IR manipulation, and PassTest registration.
Its SKILL.md is about 3.5k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
The repository describes itself as: High-Performance Rendering Framework on Stream Architectures. The licence is Apache-2.0.
6 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit 0c84a1a. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
No scripts in the folder and no shell commands in SKILL.md (its code samples are cpp and cmake).
From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Spv Opt loads about 3.5k tokens when it runs. Until then it costs about 23 tokens; SKILL.md has 503 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from LuisaGroup/LuisaCompute at commit 0c84a1a, republished under its Apache-2.0 licence (© LuisaGroup). 503 words, ~3,539 tokens.
.claude/skills/spv-opt/SKILL.md (or your agent's skills folder).Writing optimizer passes for SPIRV-Tools (src/ext/SPIRV-Tools).
Derive from spvtools::opt::Pass (or MemPass for mem2reg-style passes), implement name() and Process().
// source/opt/my_pass.h
#ifndef SOURCE_OPT_MY_PASS_H_
#define SOURCE_OPT_MY_PASS_H_
#include "source/opt/pass.h"
namespace spvtools::opt {
class MyPass : public Pass {
public:
const char* name() const override { return "my-pass"; }
Status Process() override;
IRContext::Analysis GetPreservedAnalyses() override {
return IRContext::kAnalysisDefUse | IRContext::kAnalysisCFG;
}
};
} // namespace spvtools::opt
#endif// source/opt/my_pass.cpp
#include "source/opt/my_pass.h"
namespace spvtools::opt {
Pass::Status MyPass::Process() {
bool modified = false;
// Iterate functions
for (auto& func : *get_module()) {
for (auto& block : func) {
for (auto& inst : block) {
// transform inst...
modified = true;
}
}
}
return modified ? Status::SuccessWithChange : Status::SuccessWithoutChange;
}
} // namespace spvtools::optRules:
name() must match the --my-pass CLI flag used in RegisterPassFromFlag (no leading hyphens).Process() must return Status::Failure only on real errors.Status::SuccessWithChange; the pass manager invalidates analyses not listed in GetPreservedAnalyses().Many load/store elimination passes derive from MemPass instead of Pass:
#include "source/opt/mem_pass.h"
class MyMemPass : public MemPass {
// Inherits helpers: GetPtr(), IsTargetVar(), CollectTargetVars(),
// HasOnlyNamesAndDecorates(), KillAllInsts(), Type2Undef(), etc.
};Module* m = get_module(); // or context()->module()
IRContext* ctx = context();
m->ForEachInst([](Instruction* inst){ /* all insts */ }, true);
ctx->get_def_use_mgr(); // analysis::DefUseManager
ctx->get_type_mgr(); // analysis::TypeManager
ctx->get_constant_mgr(); // analysis::ConstantManager
ctx->get_decoration_mgr(); // analysis::DecorationManager
ctx->cfg(); // CFG
ctx->GetValueNumberTable(); // ValueNumberTable
ctx->GetStructuredCFGAnalysis(); // StructuredCFGAnalysis
ctx->InvalidateAnalyses(IRContext::kAnalysisDefUse | IRContext::kAnalysisCFG);
ctx->IsConsistent(); // debug invariant checkspv::Op opcode = inst->opcode();
uint32_t rid = inst->result_id(); // 0 if none
uint32_t tid = inst->type_id(); // 0 if none
// Operands: inst->begin() .. inst->end()
for (auto& op : *inst) {
if (spvIsIdType(op.type)) { uint32_t id = op.words[0]; }
}
uint32_t val = inst->GetSingleWordInOperand(idx);
inst->NumInOperands();
inst->SetResultId(new_id);
inst->SetResultType(new_type_id);
inst->SetInOperand(idx, {new_val});
inst->ToBinary(&words);
// Predicates
inst->IsBranch();
inst->IsBlockTerminator();
inst->IsDecoration();
inst->IsConstant();
inst->IsLoad();
inst->IsNop();
inst->ToNop(); // turns instruction into OpNopfor (auto& block : func) { // func is a Function&
uint32_t label = block.id();
Instruction* label_inst = block.GetLabelInst();
Instruction* merge = block.GetMergeInst(); // OpSelectionMerge / OpLoopMerge
Instruction* loop_merge = block.GetLoopMergeInst();
bool has_phi = block.HasPhiInstructions();
// Iterate instructions (label included if you use ForEachInst)
for (auto& inst : block) { }
block.ForEachInst([](Instruction* i){ }, true); // true = include debug lines
// Terminator helpers
block.ForEachSuccessorLabels([](uint32_t id){ });
bool is_loop_header = block.IsLoopHeader();
uint32_t merge_id = block.MergeBlockIdIfAny();
uint32_t continue_id = block.ContinueBlockIdIfAny();
Instruction* term = block.terminator();
}for (auto& func : *get_module()) {
uint32_t func_id = func->DefInst().result_id();
bool is_declaration = func->IsDeclaration();
func->ForEachParam([](Instruction* param){ });
// iterate blocks (func is a Function& from the module loop)
for (auto& block : func) { }
}Function does not have an IsEntryPoint() method. Check entry points via the module:
bool IsEntryPoint(Function* func, Module* module) {
for (auto& entry : module->entry_points()) {
// OpEntryPoint: operand 0 = execution model, operand 1 = function id
if (entry.GetSingleWordInOperand(1) == func->result_id()) return true;
}
return false;
}#include "source/opt/ir_builder.h"
// Insert before an instruction
InstructionBuilder b(context(), insertion_point,
IRContext::kAnalysisInstrToBlockMapping | IRContext::kAnalysisDefUse);
// Or append to end of a block
InstructionBuilder b(context(), parent_block,
IRContext::kAnalysisInstrToBlockMapping | IRContext::kAnalysisDefUse);
Instruction* add = b.AddBinaryOp(type_id, spv::Op::OpIAdd, lhs, rhs);
Instruction* extract = b.AddCompositeExtract(elem_type_id, composite_id, {idx0, idx1});
Instruction* construct = b.AddCompositeConstruct(type_id, {id0, id1});
Instruction* load = b.AddLoad(type_id, ptr_id);
Instruction* store = b.AddStore(ptr_id, value_id);
Instruction* branch = b.AddBranch(target_id);
Instruction* cbranch = b.AddConditionalBranch(cond_id, true_id, false_id);
Instruction* cbranch_with_merge = b.AddConditionalBranch(cond_id, true_id, false_id, merge_id);
Instruction* phi = b.AddPhi(type_id, {val0, block0, val1, block1});
Instruction* unary = b.AddUnaryOp(type_id, spv::Op::OpConvertFToS, operand);
Instruction* nullary = b.AddNullaryOp(type_id, spv::Op::OpGroupAll);
Instruction* select = b.AddSelect(type_id, cond_id, true_id, false_id);
Instruction* access = b.AddAccessChain(ptr_type_id, base_ptr_id, {idx_id0, idx_id1});
Instruction* var = b.AddVariable(ptr_type_id, static_cast<uint32_t>(spv::StorageClass::Function));InstructionBuilder can only preserve kAnalysisDefUse and kAnalysisInstrToBlockMapping; other analyses must be invalidated/rebuilt explicitly.
// Replace all uses of old_id with new_id
ctx->ReplaceAllUsesWith(old_id, new_id);
// Replace uses only when predicate returns true
ctx->ReplaceAllUsesWithPredicate(old_id, new_id, [](Instruction* user) {
return user->opcode() == spv::Op::OpStore;
});
// Kill an instruction (removes from block, updates def-use)
ctx->KillInst(inst);
// Kill an id and all its uses
ctx->KillDef(id);
// Get the defining instruction for an id
Instruction* def = ctx->get_def_use_mgr()->GetDef(id);
// Get users of an id
ctx->get_def_use_mgr()->ForEachUser(id, [](Instruction* user){ });
uint32_t n = ctx->get_def_use_mgr()->NumUsers(id);When deleting instructions while iterating, collect first and kill after:
std::vector<Instruction*> to_kill;
get_module()->ForEachInst([&](Instruction* inst) {
if (inst->IsNop()) to_kill.push_back(inst);
}, false);
for (auto* inst : to_kill) context()->KillInst(inst);Phi operands arrive as (value_id, parent_block_id) pairs:
if (inst->opcode() == spv::Op::OpPhi) {
for (uint32_t i = 0; i + 1 < inst->NumInOperands(); i += 2) {
uint32_t value = inst->GetSingleWordInOperand(i);
uint32_t parent = inst->GetSingleWordInOperand(i + 1);
}
}Tests live in test/opt/. Use PassTest<::testing::Test> fixture.
#include "gmock/gmock.h"
#include "gtest/gtest.h"
#include "test/opt/pass_fixture.h"
#include "test/opt/pass_utils.h"
#include "source/opt/my_pass.h"
using MyPassTest = PassTest<::testing::Test>;
TEST_F(MyPassTest, Basic) {
const std::string before = R"(
OpCapability Shader
OpMemoryModel Logical GLSL450
%void = OpTypeVoid
%main = OpFunction %void None %void
%entry = OpLabel
OpReturn
OpFunctionEnd
)";
const std::string after = before;
SinglePassRunAndCheck<MyPass>(before, after, /*skip_nop=*/false, /*do_validation=*/false);
}Fixture helpers:
SinglePassRunAndCheck<PassT>(before, after, skip_nop, do_validation, args...) — exact match.SinglePassRunAndCheck<PassT>(before, after, skip_nop, args...) — overload without validation.SinglePassRunAndMatch<PassT>(original, do_validation, args...) — runs pass, disassembles, then checks with Effcee CHECK: patterns embedded in original. Always skips OpNop. Returns std::tuple<std::string, Pass::Status>.SinglePassRunAndFail<PassT>(original, args...) — expects Status::Failure, checks error messages with Effcee CHECK: patterns embedded in original.SinglePassRunToBinary<PassT>(assembly, skip_nop, args...) — returns std::tuple<std::vector<uint32_t>, Pass::Status>.SetAssembleOptions(SPV_TEXT_TO_BINARY_OPTION_PRESERVE_NUMERIC_IDS) — keep ids from assembly.SetDisassembleOptions(SPV_BINARY_TO_TEXT_OPTION_NO_HEADER) — omit SPIR-V header in output.SetTargetEnv(spv_target_env) — change target environment (default SPV_ENV_UNIVERSAL_1_3).For match tests, embed CHECK: lines as comments in the assembly string:
const std::string assembly = R"(
; CHECK: OpReturn
OpCapability Shader
OpMemoryModel Logical GLSL450
%void = OpTypeVoid
%main = OpFunction %void None %void
%entry = OpLabel
OpReturn
OpFunctionEnd
)";
SinglePassRunAndMatch<MyPass>(assembly, false);TEST_F(MyPassTest, Pipeline) {
const std::string before = R"(... )";
const std::string after = R"(... )";
AddPass<MyFirstPass>();
AddPass<MySecondPass>();
RunAndCheck(before, after);
}#include "source/opt/build_module.h"
TEST(MyPass, Manual) {
std::unique_ptr<IRContext> ctx =
BuildModule(SPV_ENV_UNIVERSAL_1_3, nullptr, assembly,
SPV_TEXT_TO_BINARY_OPTION_PRESERVE_NUMERIC_IDS);
ASSERT_NE(ctx, nullptr);
MyPass pass;
auto status = pass.Run(ctx.get());
EXPECT_EQ(status, Pass::Status::SuccessWithChange);
EXPECT_TRUE(ctx->IsConsistent());
}source/opt/passes.h (or include directly).CreateMyPassPass() factory declaration to include/spirv-tools/optimizer.hpp:Optimizer::PassToken CreateMyPassPass();source/opt/optimizer.cpp:Optimizer::PassToken CreateMyPassPass() {
return MakeUnique<Optimizer::PassToken::Impl>(MakeUnique<opt::MyPass>());
}source/opt/optimizer.cpp inside Optimizer::RegisterPassFromFlag:} else if (pass_name == "my-pass") {
RegisterPass(CreateMyPassPass());source/opt/CMakeLists.txt: my_pass.h
...
my_pass.cpptest/opt/ (e.g. my_pass_test.cpp) and list it in test/opt/CMakeLists.txt.Look at nearby passes in RegisterPassFromFlag for the exact pattern. Passes with arguments parse pass_args before calling RegisterPass(...).
Available IRContext::Analysis bits:
kAnalysisNone, kAnalysisDefUse, kAnalysisInstrToBlockMapping, kAnalysisDecorationskAnalysisCombinatorskAnalysisCFG, kAnalysisDominatorAnalysis, kAnalysisLoopAnalysiskAnalysisNameMap, kAnalysisScalarEvolution, kAnalysisRegisterPressurekAnalysisValueNumberTable, kAnalysisStructuredCFG, kAnalysisBuiltinVarIdkAnalysisIdToFuncMapping, kAnalysisConstants, kAnalysisTypeskAnalysisDebugInfo, kAnalysisLiveness, kAnalysisIdToGraphMappingAfter Process() returns SuccessWithChange, the pass manager automatically calls:
ctx->InvalidateAnalysesExceptFor(GetPreservedAnalyses());If you mutate IDs outside normal helpers (e.g. CompactIdsPass), manually invalidate kAnalysisDebugInfo and any others that become stale mid-pass.
If you need an analysis inside Process() and are not preserving it, it is usually fine to request it via ctx->get_def_use_mgr() etc.; the manager will build it on demand. Just make sure GetPreservedAnalyses() reflects what survives your transformations.
spvtools::Optimizer opt(SPV_ENV_UNIVERSAL_1_3);
opt.SetMessageConsumer([](spv_message_level_t, const char*, const spv_position_t&, const char* msg) {
std::cerr << msg << std::endl;
});
opt.RegisterPass(spvtools::CreateCompactIdsPass())
.RegisterPass(spvtools::CreateAggressiveDCEPass());
opt.Run(binary.data(), binary.size(), &optimized);Built-in recipes:
RegisterPerformancePasses() / RegisterSizePasses() / RegisterLegalizationPasses()bool preserve_interface argument.| File | Purpose |
|---|---|
source/opt/pass.h / pass.cpp | Base Pass class |
source/opt/mem_pass.h / mem_pass.cpp | MemPass base for mem2reg-style passes |
source/opt/empty_pass.h / null_pass.h | No-op passes for testing |
source/opt/ir_context.h / ir_context.cpp | IRContext, analysis management |
source/opt/module.h | Module, header, section lists |
source/opt/function.h | Function |
source/opt/basic_block.h | BasicBlock |
source/opt/instruction.h | Instruction, Operand, DebugScope |
source/opt/ir_builder.h | InstructionBuilder |
source/opt/def_use_manager.h | DefUseManager |
source/opt/type_manager.h | TypeManager |
source/opt/constants.h | ConstantManager |
source/opt/cfg.h | CFG |
source/opt/fold.h | Folder (constant folding) |
source/opt/passes.h | Unified include for all pass headers |
test/opt/pass_fixture.h | PassTest fixture |
include/spirv-tools/optimizer.hpp | Public C++ API |
source/opt/optimizer.cpp | Pass factories & CLI flag dispatch |
examples/cpp-interface/main.cpp | Example of assemble / validate / optimize / disassemble |
© LuisaGroup, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in .agents/skills/spv-opt of LuisaGroup/LuisaCompute.
Open the folder on GitHubat commit 0c84a1a
Spv Opt next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Spv Opt this skillLuisaGroup/LuisaCompute | 1.1k | — | ~3.5k | Automated safety check: Pass | Apache-2.0 | |
| SQL Optimizationgithub/awesome-copilot | 40k | 2 repos | ~2.3k | Automated safety check: Pass | MIT | |
| Agent Performance Optimizerruvnet/ruflo | 74k | 2 repos | ~3.6k | Automated safety check: Pass | MIT | |
| Database Optimizerdavila7/claude-code-templates | 32k | 8 repos | ~2.5k | Automated safety check: Pass | MIT | |
| Prompt Optimizeraffaan-m/ECC | 276k | 2 repos | ~2.4k | Automated safety check: Pass | MIT | |
| Cost Optimizeruvnet/ruflo | 74k | — | ~997 | Automated safety check: Notes | MIT |
github/awesome-copilot
Universal SQL performance optimization assistant for comprehensive query tuning, indexing strategies, and database performance analysis across all SQL databases (MySQL, PostgreSQL, SQL Server…
ruvnet/ruflo
Agent skill for performance-optimizer - invoke with $agent-performance-optimizer
davila7/claude-code-templates
Expert database optimizer specializing in modern performance tuning, query optimization, and scalable architectures.
affaan-m/ECC
分析原始提示,识别意图和差距,匹配ECC组件(技能/命令/代理/钩子),并输出一个可直接粘贴的优化提示。仅提供咨询角色——绝不自行执行任务。触发时机:当用户说“优化提示”、“改进我的提示”、“如何编写提示”、“帮我优化这个指令”或明确要求提高提示质量时。中文等效表达同样触发:“优化prompt”、“改进prompt”、“怎么写prompt”、“帮我优化这个指令”。不触发时机:当用户希望直接执行任…
ruvnet/ruflo
Analyze token usage patterns and recommend cost optimizations with estimated savings
JuliusBrussee/caveman
Turns a Caveman report-only optimization observation into one minimal code change and a paired baseline evaluation, after the operator picks which to pursue.
LuisaGroup/LuisaCompute
Show uncommitted changes and commit history via git. An agent skill from LuisaGroup/LuisaCompute.
LuisaGroup/LuisaCompute
Manual AST construction with FunctionBuilder for kernels and callables without DSL sugar.
LuisaGroup/LuisaCompute
CMake build options, custom functions, and backend patterns for LuisaCompute.
LuisaGroup/LuisaCompute
C++ naming, formatting, static analysis, and RTTI rules for LuisaCompute.
LuisaGroup/LuisaCompute
Debug crashes and test failures via stack-traces, host/device logging, and DSL buffer inspection.
LuisaGroup/LuisaCompute
HLSL code generation, StringBuilder patterns, builtin headers, and DXIL embedding.
SPIRV-Tools optimizer pass development, IR manipulation, and PassTest registration. Spv Opt is an agent skill from LuisaGroup/LuisaCompute. SPIRV-Tools optimizer pass development, IR manipulation, and PassTest registration.
Run `npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a claude-code`. Or copy the skill folder (.agents/skills/spv-opt in LuisaGroup/LuisaCompute) into .claude/skills/spv-opt in your project. Claude Code loads it when a task matches its description.
Run `npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a codex`. Or copy the skill folder (.agents/skills/spv-opt in LuisaGroup/LuisaCompute) into .agents/skills/spv-opt in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add LuisaGroup/LuisaCompute --skill spv-opt -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/spv-opt, .gemini/skills/spv-opt, .github/skills/spv-opt and .opencode/skills/spv-opt in your project.
SKILL.md names no scripts, command-line tools or credentials: Spv Opt is instructions for the agent only.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Spv Opt is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.5k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Spv Opt: SQL Optimization (github/awesome-copilot, 40k stars), Agent Performance Optimizer (ruvnet/ruflo, 74k stars), Database Optimizer (davila7/claude-code-templates, 32k stars) and Prompt Optimizer (affaan-m/ECC, 276k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
LuisaGroup (a GitHub organization) maintains it in LuisaGroup/LuisaCompute, which has 1,053 GitHub stars. The repository holds 12 skills in this directory. The repository was last updated on October 9, 2026.
Source: LuisaGroup/LuisaCompute on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.