Workflow·Databases

ClickHouse Build

Build ClickHouse with various configurations (Release, Debug, ASAN, TSAN, etc.).

You say
Install this skill Read the source first Free Written by ClickHouse · unverified publisher
Context cost
3k tokensestimated from the bundle, loaded when it triggers
Bundle
1 file · 11.8 kBtext throughout, nothing executable
Licence
Apache-2.0free to use
Last change
no release on file
Servers it uses
Noneruns standalone

What it does

Build ClickHouse with various configurations (Release, Debug, ASAN, TSAN, etc.). Use when the user wants to compile ClickHouse.

Installed, it changes the agent in these ways.

What this skill changes about the agent is not written down here yet. The listing was collected from its source, and the description is in its own SKILL.md.

Workflow

Runs a procedure end to end.

database
Filed under

Databases

The skill itself

This is the whole product. A skill is instructions the model reads, so there is nothing behind the listing you cannot see first — the front matter loads with every session, and the body below it loads when the skill triggers.

SKILL.md11.8 kB · 218 lines
--- name: build description: Build ClickHouse with various configurations (Release, Debug, ASAN, TSAN, etc.). Use when the user wants to compile ClickHouse. argument-hint: [build-type] [target] [options] disable-model-invocation: false allowed-tools: Task, Bash(ninja:*), Bash(cd:*), Bash(ls:*), Bash(pgrep:*), Bash(ps:*), Bash(pkill:*), Bash(mktemp:*), Bash(sleep:*) ---
9# ClickHouse Build Skill
10
11Build ClickHouse in build or build_${buildType} or build/${buildType} directory (e.g., build, build_debug, build_asan...)
12
13## Arguments
14
15- $0 (optional): Build type - one of: Release, Debug, RelWithDebInfo, ASAN, TSAN, MSAN, UBSAN. Default: RelWithDebInfo
16- $1 (optional): Target - specific target to build (e.g., clickhouse-server, clickhouse-client, clickhouse). Default: clickhouse
17- $2+ (optional): Additional cmake/ninja options
18
19## Build Types
20
21| Type | Description | CMake Flags |
22|------|-------------|-------------|
23| Release | Optimized production build | -DCMAKE_BUILD_TYPE=Release |
24| Debug | Debug build with symbols | -DCMAKE_BUILD_TYPE=Debug |
25| RelWithDebInfo | Optimized with debug info | -DCMAKE_BUILD_TYPE=RelWithDebInfo |
26| ASAN | AddressSanitizer (memory errors) | -DCMAKE_BUILD_TYPE=Debug -DSANITIZE=address |
27| TSAN | ThreadSanitizer (data races) | -DCMAKE_BUILD_TYPE=Debug -DSANITIZE=thread |
28| MSAN | MemorySanitizer (uninitialized reads) | -DCMAKE_BUILD_TYPE=Debug -DSANITIZE=memory |
29| UBSAN | UndefinedBehaviorSanitizer | -DCMAKE_BUILD_TYPE=Debug -DSANITIZE=undefined |
30
31## Common Targets
32
33- clickhouse - Main all-in-one binary
34- clickhouse-server - Server component
35- clickhouse-client - Client component
36- clickhouse-local - Local query processor
37- clickhouse-benchmark - Benchmarking tool
38
39## Build Process
40
41**IMPORTANT:** This skill assumes the build directory is already configured with CMake. Do NOT run cmake or create build directories - only run ninja.
42
431. **Determine build configuration:**
44 - Build type: $0 or RelWithDebInfo if not specified
45 - Target: $1 or clickhouse if not specified
46 - Build directory: build/${buildType} (e.g., build/RelWithDebInfo, build/Debug, build/ASAN)
47
482. **Create log file and start the build:**
49
50 **Step 2a: Create temporary log file first:**
51 ```bash
52 mktemp /tmp/build_clickhouse_XXXXXX.log
53 ```
54 - This will print the log file path
55 - **IMMEDIATELY report to the user:**
56 - "Build logs will be written to: [log file path]"
57 - Then display in a copyable code block:
58 ```bash
59 tail -f [log file path]
60 ```
61 - Example: "You can monitor the build in real-time with:" followed by the tail command in a code block
62
63 **Step 2b: Start the ninja build:**
64 ```bash
65 cd build/${buildType} && ninja [target] > [log file path] 2>&1
66 ```
67
68 **Important:**
69 - Do NOT create build directories or run cmake configuration
70 - The build directory must already exist and be configured
71 - Use the log file path from step 2a
72 - Redirect both stdout and stderr to the log file using > "$logfile" 2>&1
73 - Run the build in the background using run_in_background: true
74 - **After starting the build**, report: "Build started in the background. Waiting for completion..."
75
763. **Wait for build completion:**
77 - Use TaskOutput with block=true to wait for the background task to finish
78 - The log file path is already known from step 2a
79 - Pass the log file path to the Task agent in step 4
80
814. **Report results:**
82
83 **ALWAYS use Task tool to analyze results** (both success and failure):
84 - Use Task tool with subagent_type=general-purpose to analyze the build output
85 - **Pass the log file path from step 2a** to the Task agent - let it read the file directly
86 - Example Task prompt: "Read and analyze the build output from: /tmp/build_clickhouse_abc123.log"
87 - The Task agent should read the file and provide:
88
89 **If build succeeds:**
90 - Confirm build completed successfully
91 - Report binary location: build/${buildType}/programs/[target]
92 - Mention any warnings if present
93 - Report build time if available
94 - Keep response concise
95
96 **If build fails:**
97 - Parse the build output to identify what failed (compilation, linking, etc.)
98 - Extract and highlight the specific error messages with file paths and line numbers
99 - Identify compilation errors with the exact error text
100 - For linker errors, identify missing symbols or libraries
101 - For template errors, simplify and extract the core issue from verbose C++ template error messages
102 - Provide the root cause if identifiable
103 - Provide a concise summary with:
104 - What component/file failed to build
105 - The specific error type (syntax error, undefined symbol, etc.)
106 - File path and line number where error occurred
107 - The actual error message
108 - Brief explanation of likely cause if identifiable
109 - Filter out excessive verbose compiler output and focus on the actual errors
110
111 - Return ONLY the Task agent's summary to the user
112 - Do NOT return the full raw build output
113
114 **After receiving the summary:**
115 - If build succeeded: Proceed to step 5 to check for running server
116 - If build failed:
117 - Present the summary to the user first
118 - **MANDATORY:** Use AskUserQuestion to prompt: "Do you want deeper analysis of this build failure?"
119 - Option 1: "Yes, investigate further" - Description: "Launch a subagent to investigate the root cause across the codebase"
120 - Option 2: "No, I'll fix it myself" - Description: "Skip deeper analysis and proceed without investigation"
121 - If user chooses "Yes, investigate further":
122 - **CRITICAL: DO NOT read files, edit code, or fix the issue yourself**
123 - **MANDATORY: Use Task tool to launch a subagent for deep analysis only (NO FIXES)**
124 - Use Task tool with subagent_type=Explore to search for related code patterns, similar errors, or find where symbols/functions are defined
125 - For complex errors involving multiple files or dependencies, use Task tool with subagent_type=general-purpose to investigate missing symbols, headers, or dependencies
126 - Provide specific prompts to the agent based on the error type:
127 - Compilation errors: "Analyze the compilation error in [file:line]. Find where [symbol/class/function] is defined in the codebase and explain the root cause. Do NOT fix the code."
128 - Linker errors: "Investigate why [symbol] is undefined and find its implementation. Explain what's causing the linker error. Do NOT fix the code."
129 - Header errors: "Find which header file provides [missing declaration] and explain what's missing. Do NOT fix the code."
130 - Template errors: "Investigate the template instantiation issue with [template name] and explain the root cause. Do NOT fix the code."
131 - The subagent should only investigate and analyze, NOT edit or fix code
132 - **CRITICAL: Return ONLY the agent's summary of findings to the user**
133 - **DO NOT return full investigation details, raw file contents, or excessive verbose output**
134 - **Present findings in a well-organized summary format**
135 - If user chooses "No, I'll fix it myself":
136 - Skip deeper analysis
137 - Skip step 5 (no need to check for running server if build failed)
138
1395. **MANDATORY: Check for running server and offer to stop it (only after successful build):**
140
141 **IMPORTANT:** This step MUST be performed after every successful build. Do not skip this step.
142
143 **Use Task tool with subagent_type=general-purpose to handle server checking and stopping:**
144
145 ```
146 Task tool with subagent_type=general-purpose
147 Prompt: "Check if a ClickHouse server is currently running and handle it.
148
149 Steps:
150 1. Check for running ClickHouse server:
151 pgrep -f \"clickhouse[- ]server\" | xargs -I {} ps -p {} -o pid,cmd --no-headers 2>/dev/null | grep -v \"cmake|ninja|Building\"
152
153 2. If a server is running:
154 - Report the PID and explain it's using the old binary
155 - Use AskUserQuestion to ask: \"A ClickHouse server is currently running. Do you want to stop it so the new build can be used?\"
156 - Option 1: \"Yes, stop the server\" - Description: \"Stop the running server (you'll need to start it manually later)\"
157 - Option 2: \"No, keep it running\" - Description: \"Keep the old server running (won't use the new build)\"
158 - If user chooses \"Yes, stop the server\":
159 - Run: pkill -f \"clickhouse[- ]server\"
160 - Wait 1 second: sleep 1
161 - Verify stopped: pgrep -f \"clickhouse[- ]server\" should return nothing
162 - Report: \"Server stopped. To start the new version, run: ./build/${buildType}/programs/clickhouse server --config-file ./programs/server/config.xml\"
163 - If user chooses \"No, keep it running\":
164 - Report: \"Server remains running with the old binary. You'll need to manually restart it to use the new build.\"
165
166 3. If no server is running:
167 - Report: \"No ClickHouse server is currently running.\"
168
169 Keep the response concise and only report the outcome to the user."
170 ```
171
172 - Wait for the Task agent to complete
173 - Return the Task agent's summary to the user
174
1756. **MANDATORY: Provide final summary to user:**
176
177 After completing all steps, always provide a concise final summary to the user:
178
179 **For successful builds:**
180 - Confirm the build completed successfully
181 - Report the binary location: build/${buildType}/programs/[target]
182 - Report the server status outcome from step 5
183
184 **For failed builds:**
185 - Already handled in step 4 with error analysis and optional investigation
186
187 **Example final summary for successful build:**
188 ```
189 Build completed successfully!
190
191 Binary: build/RelWithDebInfo/programs/clickhouse
192 Server status: No ClickHouse server is currently running.
193 ```
194
195 Keep the summary brief and clear.
196
197## Examples
198
199- /build - Build clickhouse target in RelWithDebInfo mode (default)
200- /build Debug clickhouse-server - Build server in Debug mode
201- /build ASAN - Build with AddressSanitizer
202- /build Release clickhouse-client - Build only the client in Release mode
203
204## Notes
205
206- Always run from repository root
207- **NEVER** create build directories or run cmake - the build directory must already be configured
208- Build directories follow pattern: build/${buildType} (e.g., build/Debug, build/ASAN)
209- Binaries are located in: build/${buildType}/programs/
210- This skill only runs incremental builds with ninja
211- To configure a new build directory, the user must manually run CMake first
212- For a clean build, the user should remove build/${buildType} and reconfigure manually
213- **MANDATORY:** After successful builds, this skill MUST check for running ClickHouse servers and ask the user if they want to stop them to use the new build
214- **MANDATORY:** ALL build output (success or failure) MUST be analyzed by a Task agent with subagent_type=general-purpose
215- **MANDATORY:** ALWAYS provide a final summary to the user at the end of the skill execution (step 6)
216- **CRITICAL:** Build output is redirected to a unique log file created with mktemp. The log file path is reported to the user in a copyable format BEFORE starting the build, allowing real-time monitoring with tail -f. The log file path is saved from step 2a and passed to the Task agent for analysis. This keeps large build logs out of the main context.
217- **Subagents available:** Task tool is used to analyze all build output (by reading from output file) and provide concise summaries. Additional agents (Explore or general-purpose) can be used for deeper investigation of complex build errors
218
In the file
SKILL.md1,686 words
Files1
LicenceApache-2.0
Why you can read it

Nothing in a skill executes. The client loads the text and the model follows it, so a skill can be audited the way a runbook is — by reading it.

What it costs in context

Skills are not billed by the call. They are paid for in context: every token the instructions occupy is a token your code, your diff and your conversation cannot use. Here is what this one takes and when it takes it.

≈90
always loaded
The name and description, so the model knows the skill exists and when to reach for it.
2,860
on trigger
The instruction body, read only when the skill fires.
1.5%
of a 200k window
Ten skills this size would take about 15% of the window before you open a file.
050k100k150k200k context window

3k tokens, estimated from the bundle at four bytes to the token, held for the rest of the session once it triggers. Middling. Fine to keep on in a project where you use it weekly, worth unloading in one where you never do.

Servers bill, skills cost

A server charges by the month. A skill charges once per session, in context, and then keeps charging it for as long as the session lives.

Before and after

The same question, put to the same model twice: once as it comes, and once with these instructions loaded.

No worked example has been published for this skill yet.

Adoption
Installsnone yet
Ratingno reviews yet

The procedure it runs

The procedure has not been published here. It is in the skill’s own SKILL.md, which its author has not sent to the marketplace yet.

Prose, not code

These steps are written for a model to follow, not executed by a runtime. It can still be told to skip one, and it will say so when it does.

Servers it uses

None. This skill calls no MCP servers at all.

Everything it needs is in the instructions, so it works in a project with nothing connected — the model reads the file and changes how it works with what it can already reach.

It writes no files and reaches no network. All it changes is how the model reasons and writes.

What it asks for
Writes filesno
Network accessno

Read from the allowed-tools line of this skill’s own SKILL.md. A skill grants no permissions of its own — it can only ask for tools your client already has.

What it will not do

Every skill is narrow, and the useful ones say where they stop. These are the jobs this one is the wrong tool for.

What this skill is not for has not been published here. Nothing is implied by that: it is a section the author has not filled in.

What is in the bundle

1 file, 11.8 kB on disk. A bundle is text throughout: the instructions the model reads, plus the templates it fills in.

  • SKILL.md11.8 kB
What is not in it

No dependencies and nothing executable: a skill is text the agent reads, so the bundle is 1 file you can review in full before installing. The Apache-2.0 licence covers the templates and examples as well as the instructions.

Install

Installing copies the bundle into your project. Nothing runs at install time — the files sit on disk until the model reads them.

# ClickHouse Build · 3k tokens when loaded npx mcprush@latest skill add clickhouse/clickhouse-build

Writes to .claude/skills/clickhouse-build/ in the current project. Add --global to put it in your home directory instead, for every project.

Which clients pick it up on their own

A skill is a folder of text. A client with a skills folder reads it without being told; everywhere else the same text works, it is just handed to the model rather than found.

Claude Code.claude/skills/
Claude Desktop
ChatGPT
Cursor.cursor/skills/
VS Code.github/skills/
Codex CLI.agents/skills/
Gemini CLI.gemini/skills/
Grok.grok/skills/
Zed.agents/skills/
Windsurf.windsurf/skills/
Agent SDK.claude/skills/
HTTP / API
This release
Versionnot versioned
Publishedno release date on file
PriceFree
Referenceclickhouse/clickhouse-build

Versions

Its author publishes no version number, so there is nothing here to pin to: what you install is the folder as it stands today. Instructions change more often than APIs do — a skill can be rewritten entirely without anything it depends on moving.

v
  • No earlier releases have been published to the marketplace.
Pinning

Nothing to pin to: this skill carries no version number of its own. What you install is what the folder holds on the day you install it.

Reviews

no reviews yet · no installs yet

Nobody has reviewed this skill yet. The rating is the mean of the reviews written here, so there is none until somebody writes the first.

Who can post

Only accounts that have had the skill installed for fourteen days, so a review is written after living with it rather than after reading it. Publishers may reply once.

Who wrote it

CL
ClickHouse

Publishes on mcprush.

0 servers listed1 skill listednot claimed
Profile
Publisher
Servers0
Claim this skill