- Home
- Skills
- Workflow Automation
- mcplite executor
Works with the AI Agents you already use
mcplite executor
Generate JSON-based tool call batches for local file operations and bash execution via McpLite.
Free
mcplite executor
Example session with this skill installed
Find all .js files in the src folder, read the content of index.js, and then run npm test.
- Read your context and instructions
- Compiled the mcplite executor
I have prepared the tool calls to inspect your source files and execute the test suite. Paste this array into your mcplite-tasks.json file and run mcplite run.
Tool Call Batch
[
{
"id": "t1",
"tool": "glob",
"args": {
"pattern": "src/**/*.js"
},
"label": "list source files"
},
{
"id": "t2",
"tool": "read",
"args": {
"path": "src/index.js"
},
"label": "read entrypoint"
},
{
"id": "t3",
"tool": "bash",
"args": {
"command": "npm test"
},
"label": "run test suite"
}
]
Next steps
- Copy the JSON block above into
mcplite-tasks.jsonin your project root. - Run
mcplite runin your terminal. - Provide the contents of
mcplite-tasks-logs.jsonif any step fails or if you need further changes.
Connects securely to your tools. The creator never sees your data.
About this skill
Most LLMs cannot directly interact with your local filesystem or shell without a complex setup. This skill bridges that gap by generating standardized tool-call batches that you can run locally. It eliminates the friction of manual file editing and command execution by treating your project root as a programmable environment.
What it does
- Batch execution generates a single JSON array of tool calls including bash, read, write, and edit.
- File operations handles reading specific line ranges, creating directories, and performing exact string replacements.
- Pattern searching uses globbing and regex-based grep to locate files and specific code patterns across your project.
- Shell integration runs bash commands with configurable timeouts and working directories.
- Error handling provides unique task IDs and labels to correlate results and troubleshoot failed batches.
How it works
- Discovery the agent uses
globorreadto understand your project structure. - Drafting the agent emits a fenced JSON code block containing the sequence of tool calls.
- Execution you paste the JSON into
mcplite-tasks.jsonand run themcpliteCLI. - Feedback the agent analyzes the resulting logs to verify success or plan the next set of operations.
Frameworks & tools
This skill requires Node.js 18 or newer and the mcplite CLI installed globally. It works with any language or framework as long as the CLI has access to the project root.
Why this beats prompting it yourself
Writing manual instructions for file edits often leads to hallucinations or incorrect paths. This skill enforces a strict output contract that the mcplite runner understands, ensuring file operations are precise and shell commands are structured for batch processing.
Use cases
- Refactor code by finding patterns with
grepand applyingeditcalls. - Automate boilerplate creation by generating
writecalls for multiple files. - Run CI-style checks locally by chaining
bashbuild commands after file edits. - Audit large codebases by batching
globandreadoperations for analysis.
Known limitations
Requires manual copy-pasting of JSON to a local file. Does not support absolute paths or parent directory escapes. Sequential execution stops on the first thrown error by default.
Requires Node.js 18+ and global mcplite CLI. No support for absolute paths or path escapes. Sequential execution stops on error. Requires manual JSON pasting to run commands.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 11 days ago
- Free with an Agensi account
Requires