codebase-explorer
Skill3 starsChanged 10 months agoMITView on GitHub
Deep codebase exploration using semantic search and relationship mapping. Use when you need to understand the current codebase.
9 other projects use a version of this.
Adopted by
Most recently taken up by
- marcjaner/skillstheir own version, first seen Oct 20, 2026
- 0King0P/Claude-Skillstheir own version, first seen Oct 13, 2026
- NachikethReddyY/skillstheir own version, first seen Oct 13, 2026
- yashs33244/explorertheir own version, first seen Oct 13, 2026
- amirbiron/claude-skillstheir own version, first seen Oct 11, 2026
- cellrishi-code/claude-skills-atlastheir own version, first seen Oct 4, 2026
Used with
Projects with it also use code-review.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
GoogleSkillcode-review
Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.
Also in 3 projects that use this
Other versions9 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 9
- TashanGKD/tashan-cursor-skills★ 21same name
- cellrishi-code/claude-skills-atlas★ 5same name
- yashs33244/explorer★ 4same name
- marcjaner/skills★ 1same name
- mcabreradev/hermes-sdd-agency★ 1same name
- NachikethReddyY/skills★ 1same name
- srikanth-manivannan/claude-agent-factory★ 1same name
- 0King0P/Claude-Skills★ 0same name
- amirbiron/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- .claude/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- add-languageSkill
- releaseSkill
- skill-developerSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 6 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 8 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 27
Use the Bash tool to perform semantic code search:
- Distinctive: line 42
Use the line range from the Location field to read just the relevant code
- Distinctive: line 44
Use the Read tool with:
- Distinctive: line 52
If a relationship looks relevant to answering the query, investigate it
- Very specific: line 78
Add lang:rust (or python, typescript, etc.) to narrow results by language if you work on multi-language projects
- Distinctive: line 79
Follow relationships that appear in multiple results (they're likely important)
- Very specific: line 80
Use the describe command to get full details about interesting relationships
- Distinctive: line 92
Don't start implementing or making changes yet
“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Search Query Analysis
Query Optimization Skill
Codanna's semantic search works best with technical terms and specific concepts. Analyze the situation and optimize your codebase explore queries for code search:
Examples:
- If vague (e.g., "that parsing thing") → Make it specific (e.g., "language parser implementation")
- If a question (e.g., "how does parsing work?") → Extract keywords (e.g., "parsing implementation process")
- If conversational (e.g., "the stuff that handles languages") → Use technical terms (e.g., "language handler processor")
- If too broad (e.g., "errors") → Add context (e.g., "error handling exception management")
OptimizedQuery: {Claude: I will write my optimized query here, then use it below}
Execute this command with your optimized query:
Your Workflow <Workflow>
Gather Context <Step_1 GatherContext>
Use the Bash tool to perform semantic code search:
Execute: codanna mcp semantic_search_with_context query:"$OptimizedQuery" limit:5
What Codanna returns:
- Relevance scores (how well each result matches)
- Symbol signatures and documentation
- Relationships (calls, called_by, implements, defines)
- File locations with line ranges
Your Workflow <Step_2 YourWorkflow>
- Analyze the results with their relevance scores (focus on results with score > 0.6 (if possible))
- To see actual implementation of interesting results:
- Use the line range from the Location field to read just the relevant code
- Example: If you see "at
src/io/exit_code.rs:108-120" - Use the Read tool with:
file_path:src/io/exit_code.rs(use the working directory from your environment context<env>to construct the absolute path)offset: 108 (start line)limit: 13 (calculated as: 120 - 108 + 1)
- Formula:
limit = end_line - start_line + 1 - Example:
Read(file_path="/full/path/to/src/io/exit_code.rs", offset=108, limit=13)
- When relationships are shown (called_by, calls, defines, implements):
- If a relationship looks relevant to answering the query, investigate it
- Execute:
codanna retrieve describe <relationship_symbol_name|symbol_id:ID> - Example: If you see "Called by:
initialize_registry [symbol_id:123]", run:codanna retrieve describe initialize_registryordescribe symbol_id:123 - Note: Following 1-2 key relationships per result is typically sufficient
- Build a complete picture by following key relationships and reading relevant code sections
- If needed, repeat <Step_1: GatherContext> with a refined query based on what you learned.
Tips for Efficient Exploration
The results include:
- Relevance scores (how well each result matches the query)
- Symbol documentation and signatures
- Relationships (who calls this, what it calls, what it defines)
- System guidance for follow-up investigation
sed (native on unix only):
- You can also see actual implementation with
sed: (works native on Unix based environments):- Use the line range from the Location field to read just the relevant code
- Example: If you see "Location:
src/io/exit_code.rs:108-120" - Execute:
sed -n '108,120p' src/io/exit_code.rsto read lines 108-120 - This shows the actual code implementation, not just the signature. It works like the Read tool.
- Add
lang:rust(or python, typescript, etc.) to narrow results by language if you work on multi-language projects - Follow relationships that appear in multiple results (they're likely important)
- Use the
describecommand to get full details about interesting relationships
Token awareness:
- Each search uses ~500 tokens
- Each relationship follow uses ~300 tokens
- Each file read uses ~100-500 tokens (depends on size)
- Staying efficient keeps your context window clean for deeper analysis
This command is for exploration:
- Build understanding of the codebase
- Identify patterns and integration points
- Present findings and await user direction
- Don't start implementing or making changes yet
Based on the gathered context, engage with the user to narrow focus and help the user with further request.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

