Question 130
A task is to update one exact configuration value, and the file contents are already known. What tool choice minimizes unwanted changes?
- Use web fetch to inspect an unrelated public template
- Report completion after printing the proposed line to the console
- Apply a precise file edit and check the resulting diff
- Regenerate the entire configuration from a high-level guess
Explanation
Correct Answer
C:A targeted replacement limits the mutation footprint. Verification ensures the correct occurrence changed and no unrelated content moved.
Incorrect Answers
A:An external template may not match local constraints. The requested change concerns a known repository file.
B:Printing a suggestion is not a file modification. The on-disk artifact must be updated and checked.
D:Whole-file generation risks losing comments and settings. The existing file already provides a reliable baseline.
Question 131
A reviewer needs to compare an agent's edits to the repository baseline. What tool output is most useful?
- A full copy of every unchanged file with no comparison markers
- A source-control diff showing exact additions and deletions
- An unverified claim that all edits were minor
- A list of file extensions without changed content
Explanation
Correct Answer
B:A diff isolates the changed lines and supports independent review. It is stronger evidence than a prose description of intended changes.
Incorrect Answers
A:Raw repository dumps obscure the edit set and waste context. A diff targets the task.
C:Qualitative self-assessment omits the actual modifications. Reviewers need inspectable file changes.
D:Extensions show categories of files, not what the agent altered. They cannot support correctness review.
Question 132
A tool read fails because a requested file path does not exist. What is the next reasonable tool operation?
- Locate the file with path search before retrying
- Delete the parent directory to reset the workspace
- Pretend the file contained the expected configuration
- Immediately create an empty file with that path
Explanation
Correct Answer
A:A read failure may reflect a rename or incorrect path. A controlled search can identify the intended artifact without guessing.
Incorrect Answers
B:Destructive cleanup is unrelated to finding a file. It can remove valuable repository data.
C:An invented observation would corrupt subsequent edits. The agent needs actual repository evidence.
D:File creation could hide the missing dependency and make tests pass falsely. Locate the authoritative source first.
Question 133
A documentation assistant must inspect the contents of a specific known file. Which built-in operation should it prefer?
- Run a deployment command to discover whether the file exists
- Run an edit tool with an empty replacement to simulate reading
- Read the target file directly and restrict output to relevant sections
- Search the public Internet for a similarly named file
Explanation
Correct Answer
C:Direct reading provides current content without changing the workspace. Limiting output can reduce context overhead for a large file.
Incorrect Answers
A:Deployment is not a file-inspection mechanism. It carries operational risk unrelated to documentation.
B:Mutation-oriented tools are unnecessary for inspection and may have side effects. Reading is the appropriate capability.
D:A public copy may differ from the private repository. The target is already available locally.
Question 134
A repository analysis tool must create a list of import dependencies across many files. Which tool combination is appropriate?
- Treat the README alone as exhaustive dependency evidence
- Use direct code edits to insert new imports before inspecting existing ones
- Search for import statements, read key modules and synthesize the dependency graph
- Assume all files depend on each other and skip evidence collection
Explanation
Correct Answer
C:Read and search operations gather actual edges for analysis. The resulting graph can be validated without mutating code.
Incorrect Answers
A:Documentation can be outdated or incomplete. Source inspection is needed for the current graph.
B:Changing imports would alter the very dependency structure being measured. Discovery should be nonmutating.
D:A complete graph by default is inaccurate and not useful for impact analysis. Relationships must be grounded in code.
Question 135
A developer requests a safe replacement of a repeated identifier only within one function, not a repo-wide change. What should the agent do?
- Run a global search-and-replace across every project file
- Only print an example of the replacement without writing it
- Recreate the repository from scratch using generated source
- Find the function, then edit only its intended region
Explanation
Correct Answer
D:Targeted mutation follows the user's boundary and reduces collateral replacements. A diff verifies the final scope.
Incorrect Answers
A:A global replacement violates the narrow requirement and may alter unrelated APIs or documentation.
B:A suggestion does not complete the requested code modification. A controlled edit is necessary.
C:Reconstruction is disproportionate to one local rename and risks data loss. The existing code provides the edit context.
Question 136
A repository team wants Claude Code to follow its coding conventions across contributors without repeating the rules in every chat. Where should the shared instructions live?
- Inside the compiled application binary
- As a secret environment variable containing every coding rule
- In a repository CLAUDE.md file maintained with the project
- Only in each developer's latest conversational message
Explanation
Correct Answer
C:Project instructions supply persistent context for Claude Code when working in that repository. Version control makes the conventions reviewable and reproducible.
Incorrect Answers
A:An application artifact is not Claude Code's instruction-loading location. The assistant reads project context from relevant instruction files.
B:Environment variables are for configuration and secrets, not a readable team convention document. They hinder review and maintenance.
D:Temporary messages do not reliably carry over into new sessions or collaborators. Shared standards belong in a project file.
Question 137
A developer wants personal editing preferences across unrelated repositories, but team members must not inherit them. What CLAUDE.md placement is appropriate?
- Insert preferences into generated test output
- Hide preferences in an arbitrary source code comment
- Use personal user-level instructions, not project guidance
- Commit the preferences to every project root
Explanation
Correct Answer
C:User-scoped memory applies personal conventions while avoiding project-wide changes. Team standards can remain separate in checked-in project documentation.
Incorrect Answers
A:Test reports are observations, not durable assistant instruction sources. They will not establish consistent future behavior.
B:Random comments are neither reliable project guidance nor an appropriate location for global user preferences.
D:Repository files are shared configuration. Personal habits can create conflicts and unexpected behavior for other developers.
Question 138
A team has a large monorepo and needs frontend conventions only when working under its web application folder. What memory structure fits?
- Place scoped instructions in an appropriate nested CLAUDE.md or rule file for that path
- Put every frontend detail in a global user file for all developers
- Ask backend developers to manually ignore frontend rules in each session
- Remove all project instructions and depend on model familiarity with frameworks
Explanation
Correct Answer
A:Location-aware instructions reduce irrelevant guidance outside the frontend subtree. The team should verify loading semantics for the selected scope.
Incorrect Answers
B:User files are not the shared project policy and would affect unrelated repositories. Scoped project guidance is preferable.
C:Reliance on repeated human reminders scales poorly. Filesystem scope should capture the boundary where supported.
D:Generic knowledge cannot substitute for the repository's specific conventions and constraints.
Question 139
A CLAUDE.md file grows to thousands of lines of runbook notes and becomes difficult for Claude Code to apply. What change is most effective?
- Append more historical transcripts directly into CLAUDE.md
- Duplicate every instruction in multiple files to increase visibility
- Keep core instructions concise and link or import focused supporting files
- Replace specific instructions with the vague phrase work carefully
Explanation
Correct Answer
C:Relevant context should be accessible without loading unrelated detail into every task. Scoped documentation improves maintainability and attention.
Incorrect Answers
A:Raw conversations are noisy and can conflict with current instructions. Shared memory should contain curated actionable conventions.
B:Duplication wastes context and creates drift when one copy changes. A clear source of truth is preferable.
D:Generic advice omits commands, policies and project-specific expectations. The goal is focused detail, not a loss of actionable information.
Question 140
A project memory references a separate testing guide, but the team wants the guide edited independently. Which mechanism is useful?
- Reference the focused documentation from CLAUDE.md
- Replace test instructions with undocumented assumptions
- Encode the guide into an unrelated media asset
- Paste its complete contents into every new chat manually
Explanation
Correct Answer
A:A maintained linked file can preserve detailed test commands while leaving top-level guidance compact. Paths should remain valid in the repository.
Incorrect Answers
B:Generic inference may choose wrong commands or skip repository-specific checks. The explicit guide is valuable context.
C:Opaque artifacts are not a suitable instruction interface. The assistant needs readable, relevant documentation.
D:Manual repetition is error-prone and defeats reusable project configuration. The guide should remain an accessible file.