100% Latest & Updated Microsoft DP-800 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 350
A team uses GitHub Copilot across its SQL Database Project repository. The team selects: Define repository instructions for schema and test conventions. Which implementation action puts this design into practice?
- Create .github/copilot-instructions.md with approved coding policy
- Check tenant settings, capacity and access prerequisites
- Create dedicated least-privileged database principal for tool
- Scan repository examples and apply secret-detection hooks
Explanation
Correct Answer
A:Create .github/copilot-instructions.md with approved coding policy. This is the concrete setup for define repository instructions for schema and test conventions, and assess suggested migrations against the repository rules supplies a direct confirmation.
Incorrect Answers
B:Check tenant settings, capacity and access prerequisites. This concerns copilot in fabric for SQL development assistance. Instead, create .github/copilot-instructions.md with approved coding policy.
C:Create dedicated least-privileged database principal for tool. This is relevant when an ai assistant uses a database connection during debugging. Instead, create .github/copilot-instructions.md with approved coding policy.
D:Scan repository examples and apply secret-detection hooks. This concerns a repository contains examples that copilot may use in suggestions. Instead, create .github/copilot-instructions.md with approved coding policy.
Question 351
A team uses GitHub Copilot across its SQL Database Project repository. After implementing the design, which check most directly validates the expected behavior?
- Verify approved users can invoke assistant in the intended workspace
- Inspect tool transcripts for credential leakage
- Assess suggested migrations against the repository rules
- Verify new pull requests fail on leaked secret patterns
Explanation
Correct Answer
C:Assess suggested migrations against the repository rules. This verifies the requirement in the case rather than merely confirming that a job ran; an opposing result can reveal that instructions file is absent or not applied.
Incorrect Answers
A:Verify approved users can invoke assistant in the intended workspace. This concerns copilot in fabric for SQL development assistance. Here, assess suggested migrations against the repository rules.
B:Inspect tool transcripts for credential leakage. This concerns a prompt asks an assistant to print database passwords for troubleshooting. Here, assess suggested migrations against the repository rules.
D:Verify new pull requests fail on leaked secret patterns. This concerns a repository contains examples that copilot may use in suggestions. Here, assess suggested migrations against the repository rules.
Question 352
A team uses GitHub Copilot across its SQL Database Project repository. A test instead finds: Suggestions ignore established SQL code conventions. Which underlying issue best explains this finding?
- Instructions file is absent or not applied
- Tenant or workspace feature prerequisites are not enabled
- MCP connection scope is overly broad
- Unsafe examples were committed to accessible context
Explanation
Correct Answer
A:Instructions file is absent or not applied. It accounts for the specific observed failure: suggestions ignore established sql code conventions. The intended solution is define repository instructions for schema and test conventions with the required setup.
Incorrect Answers
B:Tenant or workspace feature prerequisites are not enabled. This concerns assistant is inaccessible in a permitted workspace. Here, instructions file is absent or not applied.
C:MCP connection scope is overly broad. This concerns agent can query an unintended workspace. Here, instructions file is absent or not applied.
D:Unsafe examples were committed to accessible context. This concerns assistant repeats a credential from an example file. Here, instructions file is absent or not applied.
Question 353
A developer invokes a third-party MCP server from the SQL code assistant. Which database design or engineering approach most directly meets this requirement?
- Restrict MCP tools and endpoints to trusted approved servers
- Define repository instructions for schema and test conventions
- Apply least-privilege security review to AI-assisted output
- Keep human review of AI-generated DDL and its data impact
Explanation
Correct Answer
A:Restrict MCP tools and endpoints to trusted approved servers. The implementation is allowlist endpoints, scope credentials and tool permissions; success can be checked by review invoked tool calls and returned resources. The method addresses the stated data and workload constraint.
Incorrect Answers
B:Define repository instructions for schema and test conventions. This is relevant when a team uses github copilot across its SQL database project repository. Instead, restrict MCP tools and endpoints to trusted approved servers.
C:Apply least-privilege security review to AI-assisted output. This concerns a code assistant proposes SQL access beyond the developer's privileges. Instead, restrict MCP tools and endpoints to trusted approved servers.
D:Keep human review of AI-generated DDL and its data impact. This concerns explain generated SQL before approving deployment. Instead, restrict MCP tools and endpoints to trusted approved servers.
Question 354
A developer invokes a third-party MCP server from the SQL code assistant. The team selects: Restrict MCP tools and endpoints to trusted approved servers. Which implementation action puts this design into practice?
- Allowlist endpoints, scope credentials and tool permissions
- Review permissions and parameterize generated queries
- Create dedicated least-privileged database principal for tool
- Scan repository examples and apply secret-detection hooks
Explanation
Correct Answer
A:Allowlist endpoints, scope credentials and tool permissions. This is the concrete setup for restrict mcp tools and endpoints to trusted approved servers, and review invoked tool calls and returned resources supplies a direct confirmation.
Incorrect Answers
B:Review permissions and parameterize generated queries. This concerns a code assistant proposes SQL access beyond the developer's privileges. Here, allowlist endpoints, scope credentials and tool permissions.
C:Create dedicated least-privileged database principal for tool. This is relevant when an ai assistant uses a database connection during debugging. Here, allowlist endpoints, scope credentials and tool permissions.
D:Scan repository examples and apply secret-detection hooks. This concerns a repository contains examples that copilot may use in suggestions. Here, allowlist endpoints, scope credentials and tool permissions.
Question 355
A developer invokes a third-party MCP server from the SQL code assistant. After implementing the design, which check most directly validates the expected behavior?
- Run with a restricted test principal and inspect access
- Validate access to allowed lakehouse metadata only
- Review invoked tool calls and returned resources
- Inspect tool transcripts for credential leakage
Explanation
Correct Answer
C:Review invoked tool calls and returned resources. This verifies the requirement in the case rather than merely confirming that a job ran; an opposing result can reveal that untrusted mcp endpoint was enabled with broad access.
Incorrect Answers
A:Run with a restricted test principal and inspect access. This concerns a code assistant proposes SQL access beyond the developer's privileges. Instead, review invoked tool calls and returned resources.
B:Validate access to allowed lakehouse metadata only. This concerns connection to a fabric lakehouse through mcp. Instead, review invoked tool calls and returned resources.
D:Inspect tool transcripts for credential leakage. This concerns a prompt asks an assistant to print database passwords for troubleshooting. Instead, review invoked tool calls and returned resources.
Question 356
A developer invokes a third-party MCP server from the SQL code assistant. A test instead finds: An unexpected tool exfiltrates schema information. Which underlying issue best explains this finding?
- Raw credentials were embedded in prompt or tool results
- MCP connection scope is overly broad
- Generated SQL was trusted without privilege checks
- Untrusted MCP endpoint was enabled with broad access
Explanation
Correct Answer
D:Untrusted MCP endpoint was enabled with broad access. It accounts for the specific observed failure: an unexpected tool exfiltrates schema information. The intended solution is restrict mcp tools and endpoints to trusted approved servers with the required setup.
Incorrect Answers
A:Raw credentials were embedded in prompt or tool results. This concerns a generated transcript contains connection string password. Here, untrusted MCP endpoint was enabled with broad access.
B:MCP connection scope is overly broad. This concerns agent can query an unintended workspace. Here, untrusted MCP endpoint was enabled with broad access.
C:Generated SQL was trusted without privilege checks. This concerns copilot-generated query reads a restricted table. Here, untrusted MCP endpoint was enabled with broad access.
Question 357
A developer must explain generated SQL before approving deployment. Which database design or engineering approach most directly meets this requirement?
- Keep human review of AI-generated DDL and its data impact
- Remove sensitive data and constrain repository access
- Supply read-only credentials for exploratory AI tool calls
- Apply least-privilege security review to AI-assisted output
Explanation
Correct Answer
A:Keep human review of AI-generated DDL and its data impact. The implementation is require a pull request with explain plans, tests and reviewers; success can be checked by inspect schema diff and execution plans prior to merge. The method addresses the stated data and workload constraint.
Incorrect Answers
B:Remove sensitive data and constrain repository access. This concerns a repository contains examples that copilot may use in suggestions. Instead, keep human review of AI-generated DDL and its data impact.
C:Supply read-only credentials for exploratory AI tool calls. This is relevant when an ai assistant uses a database connection during debugging. Instead, keep human review of AI-generated DDL and its data impact.
D:Apply least-privilege security review to AI-assisted output. This concerns a code assistant proposes SQL access beyond the developer's privileges. Instead, keep human review of AI-generated DDL and its data impact.
Question 358
A developer must explain generated SQL before approving deployment. The team selects: Keep human review of AI-generated DDL and its data impact. Which implementation action puts this design into practice?
- Use secret store and redacted placeholders in instructions
- Review permissions and parameterize generated queries
- Require a pull request with explain plans, tests and reviewers
- Allowlist endpoints, scope credentials and tool permissions
Explanation
Correct Answer
C:Require a pull request with explain plans, tests and reviewers. This is the concrete setup for keep human review of ai-generated ddl and its data impact, and inspect schema diff and execution plans prior to merge supplies a direct confirmation.
Incorrect Answers
A:Use secret store and redacted placeholders in instructions. This concerns a prompt asks an assistant to print database passwords for troubleshooting. Instead, require a pull request with explain plans, tests and reviewers.
B:Review permissions and parameterize generated queries. This concerns a code assistant proposes SQL access beyond the developer's privileges. Instead, require a pull request with explain plans, tests and reviewers.
D:Allowlist endpoints, scope credentials and tool permissions. This concerns a developer invokes a third-party mcp server from the SQL code assistant. Instead, require a pull request with explain plans, tests and reviewers.
Question 359
A developer must explain generated SQL before approving deployment. After implementing the design, which check most directly validates the expected behavior?
- Review invoked tool calls and returned resources
- Run with a restricted test principal and inspect access
- Inspect tool transcripts for credential leakage
- Inspect schema diff and execution plans prior to merge
Explanation
Correct Answer
D:Inspect schema diff and execution plans prior to merge. This verifies the requirement in the case rather than merely confirming that a job ran; an opposing result can reveal that ai output bypassed the normal code-review gate.
Incorrect Answers
A:Review invoked tool calls and returned resources. This concerns a developer invokes a third-party mcp server from the SQL code assistant. Instead, inspect schema diff and execution plans prior to merge.
B:Run with a restricted test principal and inspect access. This concerns a code assistant proposes SQL access beyond the developer's privileges. Instead, inspect schema diff and execution plans prior to merge.
C:Inspect tool transcripts for credential leakage. This concerns a prompt asks an assistant to print database passwords for troubleshooting. Instead, inspect schema diff and execution plans prior to merge.
Question 360
A developer must explain generated SQL before approving deployment. A test instead finds: Unreviewed DDL drops a dependent object. Which underlying issue best explains this finding?
- Tenant or workspace feature prerequisites are not enabled
- AI output bypassed the normal code-review gate
- Generated SQL was trusted without privilege checks
- The AI tool connection used a write-capable principal
Explanation
Correct Answer
B:AI output bypassed the normal code-review gate. It accounts for the specific observed failure: unreviewed ddl drops a dependent object. The intended solution is keep human review of ai-generated ddl and its data impact with the required setup.
Incorrect Answers
A:Tenant or workspace feature prerequisites are not enabled. This concerns assistant is inaccessible in a permitted workspace. The relevant SQL or system operation is AI output bypassed the normal code-review gate.
C:Generated SQL was trusted without privilege checks. This concerns copilot-generated query reads a restricted table. The relevant SQL or system operation is AI output bypassed the normal code-review gate.
D:The AI tool connection used a write-capable principal. This concerns agent accidentally changes production rows. The relevant SQL or system operation is AI output bypassed the normal code-review gate.
Developing AI-Enabled Database Solutions (DP-800) Practice Tests