Robust Edit

Skill: Robust Surgical Editing

This skill provides a three-step protocol to ensure that the replace tool succeeds on the first attempt by eliminating guesswork regarding whitespace and literal content.

The Protocol

Step 1: Locating the Target (Grep)

Use grep_search with a specific pattern to identify the target file and the approximate line numbers. Do not proceed until you have a confirmed file path and a line number range.

Step 2: Capturing Literal Content (Read Exact)

Once the region is identified, use read_file with the start_line and end_line parameters.

• Goal: Capture the EXACT literal string of the code you intend to replace.
• Critical: This includes all leading/trailing whitespace, tabs, and newline characters.

Step 3: Executing the Replacement (Replace)

Using the literal string captured in Step 2 as your old_string, perform the replacement.

• Instruction: In the replace tool call, provide the literal text exactly as it appeared in the read_file output.
• Validation: If the tool returns a "replacement failed" error, DO NOT RETRY with a guessed string. Repeat Step 2 to re-capture the literal content and identify any invisible characters or unexpected formatting.

Best Practices

• Minimize Scope: Keep the old_string as small as possible while remaining unique.
• Avoid Wildcards: The replace tool expects a literal match. Do not use regex or placeholders in old_string.
• Batching: You can read multiple regions in parallel before executing a batch of replacements.