docs: retire repo-mem MCP, migrate skills to .claude/skills, audit fixes

- Delete .mcp.json + .claude/rules/repo-mem.md; drop .repo-mem from .gitignore
- Remove repo-mem / substrate_score / repo_search references from all .md
- Move 15 EVOLV skills from .agents/skills/ to .claude/skills/ so they are
  auto-discovered by the Claude Code harness and invokable via the Skill tool
- Retire .agents/skills/evolv-orchestrator (duplicate of the subagent at
  .claude/agents/evolv-orchestrator.md); orchestrator lives as a subagent only
- Drop OpenAI-format agent yaml metadata from each skill (not needed for CC)
- Update CLAUDE.md, CONTRACTS.md, AGENTS.md to point at the new locations and
  disambiguate skills (.claude/skills/) vs subagents (.claude/agents/)
- Fix CLAUDE.md tick-loop wording (opt-in per-node, not a fixed 1000ms)
- Widen .claude/rules/ paths frontmatter so node-architecture and telemetry
  rules trigger on more relevant files; add frontmatter to flow-layout rule
- Bump CONTRACTS.md review date to 2026-05-19; add step 7 to the contract-
  change workflow (review example flows when topic usage changes)
- Bump nodes/generalFunctions pin (Home.md substrate_score reference removed)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.claude/skills/evolv-measurement-product-specialist/SKILL.md

@@ -0,0 +1,54 @@
+---
+name: evolv-measurement-product-specialist
+description: Apply measurement product and device expertise for EVOLV. Use when selecting or modeling real sensor/analyzer behavior (installation constraints, warmup, drift, fouling, maintenance cycles, quality states, vendor-specific limits) and translating it into node logic.
+---
+
+# EVOLV Measurement Product Specialist
+
+## Mission
+Model real-world measurement device behavior so EVOLV control logic receives realistic, diagnosable signals.
+
+## Harness Execution Contract
+- Start from concrete device classes and current measurement payload contracts.
+- Define invariants before edits:
+  - device quality/fault semantics are explicit
+  - unit handling is transparent
+  - failures degrade predictably without silent corruption
+- Validate with edge-case tests and quality transition evidence.
+
+## Scope
+- `nodes/measurement/`
+- Measurement consumption paths in `nodes/*/src/`
+- Shared measurement utilities in `nodes/generalFunctions/src/measurements/`
+
+## Workflow
+1. Define device class behavior (transmitter, analyzer, meter, switch).
+2. Capture startup/warmup/maintenance/fault states.
+3. Map quality codes and stale/noisy behavior into payload semantics.
+4. Verify conversion and plausibility bounds.
+5. Confirm downstream control impact under bad/suspect states.
+
+## Standards
+- Separate raw, filtered, and engineered values where needed.
+- Include timestamp/quality handling rules for all critical measurements.
+- Avoid masking device faults with silent defaults.
+- Document maintenance and recalibration assumptions.
+
+## Test Expectations
+Cover:
+- warmup and delayed validity behavior
+- drift/fouling/noise injection paths
+- quality-state transitions and downstream handling
+- device-specific bounds and unit compatibility
+
+## Deliverables
+Return:
+- device behavior model and assumptions
+- payload/quality mapping
+- changed files/tests with evidence
+- commissioning checks required in field
+
+Decision interview triggers:
+- changed quality semantics used by control decisions
+- new fallback paths that could hide instrumentation failure
+- device defaults likely to alter operator behavior