Context-compensation scaffold for LLM evaluation prompts. A short language prefix you prepend so the model discloses prior exposure, scores on quoted evidence only, and hedges on thin evidence — for scorers that can see your CLAUDE.md, memory, or session context. Backend-agnostic. Experimental: variance-reduction effect not yet measured.

evaluation scaffold ai-safety ai-agents rubric multi-turn debiasing llm prompt-engineering evals llm-evaluation ai-reliability lpci hermes-labs context-compensation language-as-state agent-scaffold drift-recovery recovery-scaffold

• Updated May 27, 2026
• Python

Drift-prevention session-init convention card for fresh Claude Code sessions. Injects a self-contained card so a new session opens with its grounding triggers, calibration rules, and tool-map in scope — instead of re-deriving them at minute 30. Per-project, marker-anchored, idempotent, reversible. Bash installer plus MCP server.

bash mcp developer-tools ai-safety ai-agents llm prompt-engineering agent-orchestration claude-code context-engineering claude-md ai-reliability llm-tooling drift-prevention lpci hermes-labs session-bootstrap session-init language-as-state agent-scaffold

• Updated May 28, 2026
• Python

langquant (LPCI) is a scaffold-as-state research artifact testing whether a refreshing language scaffold can serve as the sole working state for a stateless LLM. In one A/B run (n=1/condition, 20 turns) the model held coherence with zero history; transfer entropy dropped 0.608 to 0.085, a large reduction, not zero. Single observation, not a proof.

research scaffold stateless ai-agents transfer-entropy llm agent-memory context-engineering stateless-llm ai-reliability research-artifact lpci hermes-labs

• Updated Jun 7, 2026
• Python