Remove auxiliary badge guide

.bestpractices.json

@@ -67,5 +67,5 @@
   "dynamic_analysis_unsafe_justification": "The passing criterion explicitly permits N/A when the project does not produce software written in a memory-unsafe language. Maintained first-party application and tooling sources are C# without unsafe blocks or AllowUnsafeBlocks; the public source tree contains no first-party C/C++ implementation: https://github.com/codegiveness/postgresql-sharp-mcp/tree/bde167e95e0fb3ff74b895db55c87f4b04d69b0c/src and https://github.com/codegiveness/postgresql-sharp-mcp/tree/bde167e95e0fb3ff74b895db55c87f4b04d69b0c/tools . Native runtime/dependency code is not a claim of memory-safety certification.",
   "dynamic_analysis_enable_assertions_status": "Met",
   "dynamic_analysis_enable_assertions_justification": "Public-source evidence: the verification configuration runs deterministic varied-input SQL checks with always-enabled Check.Equal/Check.That and expected-error assertions, including exact PostgreSQL literal round trips, no unexpected truncation and second-statement/control-statement rejection: https://github.com/codegiveness/postgresql-sharp-mcp/blob/bde167e95e0fb3ff74b895db55c87f4b04d69b0c/tools/PostgreSqlMcp.Verify/FuzzChecks.cs . These assertions belong to the test verifier, not the production server, and integration succeeded in https://github.com/codegiveness/postgresql-sharp-mcp/actions/runs/36975816424 . This does not establish release-specific pre-release analysis or any branch-coverage percentage.",
-  "general_comments": "Repository-maintained answer proposals only; this file does not save hosted answers for https://www.bestpractices.dev/en/projects/15155 . Human-readable question text, choices and paste-ready explanations are provided in docs/best-practices-checklist.md. Default-branch policy evidence applies when this revision is published on main. No answers are proposed from subjective developer-knowledge attestations, presumed report-response histories, an absence of scanner findings, assumed credential cleanliness, unmeasured branch coverage or unverified release-specific pre-release analysis."
+  "general_comments": "Evidence-backed project metadata for https://www.bestpractices.dev/en/projects/15155 . Proposed answers require review and saving on the badge website. Developer-knowledge attestations, report-response histories, credential cleanliness, branch coverage and release-specific analysis are not inferred."
 }

CHANGELOG.md

@@ -4,7 +4,6 @@
 
 - Verify Best Practices enrollment as project 15155 (19%, not passing); add evidence-backed owner-review proposals and correct stale enrollment, secret-feature and Scorecard findings documentation without claiming hosted changes.
 - Display the live OpenSSF Best Practices badge in README, including its in-progress state; distinguish the saved owner self-assessment from local answer proposals and security certification.
-- Provide a Best Practices owner checklist using the questionnaire's exact visible question text, answer choices and copy-ready explanations instead of requiring manual interpretation of internal JSON identifiers.
 - Harden solo-maintainer release execution: require a manual main-branch dispatch and an existing tag reachable from main; compile trusted orchestration in read-only preflight and use digest-checked immutable artifacts in checkout-free attestation/publishing jobs. Replace source discovery in publishers with an explicit publication context.
 - Preserve zero required approvals while enabling resolved review conversations, enforced full-SHA Actions pins, a restricted action-publisher allowlist and owner approval for external fork workflows. Restrict the release environment to main; document unchanged optional secret controls and that hosted badge answers still require owner authentication.
 - Replace the coverage-guided SQL fuzz Bash launcher with checksum-verified, deadline-bounded C# orchestration and fail-closed seed replay; keep portable corpus names and repair the asynchronous FsCheck verifier entrypoint. Document distinct PostgreSQL property and lexer fuzzing layers.

README.md

@@ -19,8 +19,6 @@ A PostgreSQL MCP server built with C#/.NET 10 and Npgsql. One stdio server expos
 
 The server, npm installation helper, packaging, verification and release automation are C#/.NET. No maintained JavaScript, Python or Bash implementation is required. npm itself requires Node.js for installation; the installed server runs directly as a .NET executable, without a Node process. Workflow badges and Scorecard report checks and practices, not certifications or profile achievements. Best Practices shows the saved owner self-assessment, which may still be in progress. See [Security posture](docs/security-posture.md) for supply-chain evidence and limits.
 
-For the Best Practices questionnaire, use the [exact website questions and copy-ready answers](docs/best-practices-checklist.md). Repository proposals do not save the website's answers; the project owner must review and save them.
-
 ## Quick start
 
 ### 1. Install or run