- Home
- Skills
- Mhattingpete
- Claude Skills Marketplace
- Code Execution
code-execution_skill
- HTML
269
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill mhattingpete/claude-skills-marketplace --skill code-execution- SKILL.md3.1 KB
Overview
This skill executes Python code locally with integrated marketplace API access to enable large-scale code operations with dramatic token savings (typically 90%+). It is designed for bulk refactors, audits, and iterative multi-file workflows while returning concise metadata summaries instead of raw source. Use it to move heavy computation and file changes off the conversational context and keep responses compact and actionable.
How this skill works
The skill exposes a runtime API with filesystem, code analysis, transformation, and git primitives. Code analysis returns metadata only (no full source), transformations operate across file globs, and filesystem helpers copy and paste code blocks locally. Typical flow: analyze metadata, perform batch processing locally, then return a summary of results and counts rather than file contents.
When to use it
- Bulk operations affecting 10+ files (refactors, audits, migrations)
- Complex multi-step workflows that require iterative processing
- Performance-sensitive tasks where token cost matters
- Batch transformations or pattern-based renames across repos
- Generating summaries or metrics (e.g., complexity, replacements)
Best practices
- Return summaries and counts, never full source code in replies
- Use code analysis APIs for metadata only (avoid reading whole files)
- Run operations with glob patterns to batch-process files
- Handle errors and surface an error count and example cases
- Commit changes via the git API after verifying results locally
Example use cases
- Rename an identifier across a 50-file Python codebase and report files modified and total replacements
- Audit 100+ files for complexity and return a list of high-complexity filenames and counts
- Extract and move utility functions into a shared module, reporting how many functions moved
- Remove debug statements project-wide, returning number of files changed and removed statements
- Perform staged edits, run tests locally, and commit a feature branch with a concise summary
FAQ
No. The runtime returns metadata and summaries only; avoid returning raw source to keep context small and secure.
How much token savings can I expect?
Typical savings range from ~90% for small bulk jobs to over 99% for very large jobs, because work happens locally instead of in the conversational context.