test: police MCP doc contracts for every tool

[cyberlife-coder]

Closes #1695

Summary

• make structural section, source-method and list-item ownership authoritative before the proximity fallback
• police all 23 tools in the published MCP capture across documentation, skills, Node, Python, WASM and TypeScript facades
• distinguish nested item shapes, input/config envelopes and deliberate binding divergences with exact stale-checked contracts
• pin every detected declaration so deletion cannot silently reduce coverage
• add prose-enumeration coverage and one executable refusal mutation for each of the three rollout batches
• repair the return-shape drift surfaced in why, compile_context, retrieve_context_source and extraction documentation

Validation

• python3 -m unittest discover -s scripts/tests -p "test_*.py" -t . -v — 472 passed, 9 skipped (PowerShell unavailable)
• MCP guard targeted suite — 74 passed
• refusal and accepted-control vectors — passed
• python3 scripts/check-mcp-doc-contract.py — 228 surfaces, passed
• bash scripts/run-production-gates.sh — all 5 gates passed
• native Node/Python/WASM checks and Clippy with RUSTFLAGS=-D warnings — passed
• WASM no-default target check with RUSTFLAGS=-D warnings — passed
• TypeScript typecheck and Vitest — 927 passed
• cargo fmt --all -- --check, JSON/Python syntax and diff checks — passed

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 0 complexity

Metric	Results
Complexity	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}