fix(mcp): align tool schemas with accepted inputs

[ben-ranford]

Summary

Addresses unresolved Copilot review feedback from the merged MCP PR by tightening the advertised MCP tool contracts and no-argument decoding behavior.

Changes

• Advertises topN in the single-dependency MCP tool schema because the argument is accepted and ignored for dependency runs.
• Adds anyOf alternatives to the baseline comparison schema for baselinePath or baselineStorePath plus baselineKey.
• Treats JSON null params/arguments as empty objects in strict MCP decoding.

Validation

Commands and checks run:

go test ./internal/mcp
go test ./...
make lint
pre-commit make ci

Additional manual validation:

• Verified this maps directly to the three unresolved Copilot threads on PR feat(mcp): add stdio server tools #1020.

Risk and compatibility

• Breaking changes: None
• Migration required: None
• Performance impact: None expected; schema/decoder-only contract fix.
• Memory benchmark impact: None; pre-commit memory benchmark gate passed with no meaningful delta.

Checklist

• Tests added/updated for behavior changes
• Docs updated (README/docs/schema) if needed
• memory-approved requested/applied if intentional memory benchmark regressions exceed CI thresholds
• No unrelated changes included
• Ready for review

[Copilot]

Pull request overview

This PR tightens the MCP stdio tool contracts to better match actual accepted inputs and makes strict decoding more tolerant of JSON null for params/arguments, addressing the remaining review feedback from PR #1020.

Changes:

• Advertise topN as an accepted (ignored) argument for the single-dependency analysis tool schema.
• Express baseline comparison inputs as schema alternatives: baselinePath or (baselineStorePath + baselineKey).
• Treat JSON null as {} during strict MCP decoding, with added test coverage.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
internal/mcp/tools.go	Updates MCP tool input schemas and makes decodeStrict treat null as an empty object.
internal/mcp/server_test.go	Extends tools/list coverage to assert the updated schema shape (topN presence and baseline anyOf).
internal/mcp/coverage_test.go	Adds a branch test ensuring decodeStrict accepts null arguments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Memory Benchmarks

Thresholds: bytes/op <= +15.0%, allocs/op <= +10.0%

Benchmark	Base B/op	Head B/op	Delta B/op	Base allocs/op	Head allocs/op	Delta allocs/op	Status
github.com/ben-ranford/lopper/internal/lang/shared/BenchmarkCountUsage	25632.0	25632.3	+0.0%	375.0	375.0	+0.0%	ok
github.com/ben-ranford/lopper/internal/lang/shared/BenchmarkCountUsageRegexPerIdentifier	413244.7	414130.0	+0.2%	3067.0	3067.0	+0.0%	ok
github.com/ben-ranford/lopper/internal/report/BenchmarkFormatLargeTable	256581.3	256578.0	-0.0%	3079.0	3079.0	+0.0%	ok

Result: memory benchmark gate passed.

Approval: not required.

[github-actions]

Lopper (Delta)

Metric delta	Value
Dependency count	+0
Used percent	+0.0%
Waste percent	+0.0%
Estimated unused bytes	+0 B
Known licenses	+0
Unknown licenses	+0
Denied licenses	+0

Changed	Regressions	Progressions	Added	Removed	Unchanged
0	0	0	0	0	9

No dependency-surface deltas detected.

[github-actions]

SonarQube (PR)

Open issues: 0
Actionable issues shown (excluding mock/fixture files): 0

Duplication

• Overall duplicated lines: 768
• Overall duplication density: 0.60%
• New duplicated lines: n/a
• New duplication density: n/a

Issues

Open Sonar issues (0)

#	Severity	Rule	Location	Message
-	-	-	-	No open Sonar issues for this PR.

Source: SonarCloud PR view

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud