Add constraint intelligence operating layer

[leadingproblemsolver]

Summary

This implements the phase change from GitHub comment-count optimization to a constraint-first intelligence and exploitation system.

What changed

• preserves issue bodies and labels in discovery artifacts instead of throwing away the evidence needed downstream
• replaces the default discovery query set with production-constraint searches across runtime truth, side effects, identity/credentials, persistence/resume, delivery, resource pressure, configuration fidelity, and buyer/workaround cost
• adds a versioned root-constraint taxonomy
• adds build_constraint_intelligence.py to classify signals, build economic chains, score constraint value independently from intervention value, assign saturation states, and route actions
• adds compile_constraint_corpus.py to merge/deduplicate multiple signal sources before compilation
• warm-starts the corpus with data/constraint_seed_signals.json, containing explicitly tagged observed_summary records from the strongest production signals already gathered across n8n, Hermes, Agno, Mastra, PydanticAI, Paseo, and LlamaIndex
• routes occurrences through corpus_only, watch, ask_diagnostic_question, precision_comment, build_reproduction, build_patch_or_diagnostic, or operator_assistance_or_offer
• adds tests proving high-value saturated threads can remain corpus-only while unsaturated high-value signals escalate
• updates the scheduled GitHub workflow to run tests, discover fresh signals, merge them with accumulated evidence, compile the constraint leaderboard, and upload the intelligence artifacts
• documents exploration/exploitation gates, proof ladder, daily/weekly loops, public contribution gate, KPIs, and the first three audit offers

Key policy change

Default public action is no comment. A high-value market signal is preserved even when its intervention value is low.

Outputs

• artifacts/github-review-queue.json / .md
• artifacts/constraint-intelligence.json
• artifacts/constraint-leaderboard.json
• artifacts/constraint-intelligence.md

Evidence boundary

Seed records are marked as observed summaries and are not represented as verbatim issue bodies. This does not expand LCE's product claims. Constraint classification/scoring is a heuristic operational-analysis layer; LCE remains the deterministic local-first evidence compiler described in the README.

Validation

The workflow runs the discovery and constraint-intelligence tests before generating artifacts.