andrej-karpathy-skills CLAUDE md at main multica-ai andrej-karpathy-skills

7-agent multi-perspective review with quality rubrics. 12-agent paper writing pipeline. Per-agent responsibilities and per-stage artifacts now live in docs/ARCHITECTURE.md. The experiment_ids are καλαμακιου frozen at this intake point; the writers later reference them via planned_experiment_ids. If you ran experiments externally, the scholar enters one experiment_provenance entry per experiment (experiment_id, nested repro_lock, planned_vs_executed, negative_results, known_limitations) and the declaration is set to experiments_declared; if not, it is set to no_experiments_declared. If your research involves running experiments (code or human studies) before writing, the Experiment Agent skill fills the gap between ARS Stage 1 (RESEARCH) and Stage 2 (WRITE).

No new features and no breaking schema changes. Proposed reviewer-response tooling in PR #433; the deep-research three-way-scan mode and the academic-paper rebuttal-audit mode (rescued from the PR's audit concept) were integrated from that contribution in v3.12.1. It's about engaging AI deeply enough to discover its structural limits yourself — and your own thinking limits in the process. But now there are explicit checkpoints that slow down the sycophancy, force the DA to justify concessions, and prevent the Mentor from wrapping up before the user is ready. Devil's Advocate — Concession Threshold Protocol (deep-research + academic-paper-reviewer)

Imported skills carry the ARS methodology (research / writing / review protocols); Claude Code-specific machinery — slash commands, hooks, subagent orchestration — does not transfer. Then try /ars-plan to walk through your paper structure via Socratic dialogue, or jump to Quick install for prerequisites and the traditional symlink flow. /plugin marketplace add Imbad0202/academic-research-skills /plugin install academic-research-skills Merge with project-specific instructions as needed. Claude academic-writing literature-review peer-review ai-research prompt-engineering claude-code academic-pipeline V3.6.2 introduces Schema 13 sprint contracts and a hard-gate orchestration that forces reviewers to pre-commit their scoring plan before reading the paper.

Architecture & pipeline

👉 docs/PERFORMANCE.md — per-mode token budgets, full-pipeline estimate (~$4–6 for a 15k-word paper), and recommended Claude Code settings (Auto mode; Agent Team optional). See docs/SETUP.md Method 5 for details. The architecture doc supersedes the sprawling pipeline description that used to live here.

Companion: Experiment Agent

• Socratic mode (deep-research) and Plan mode (academic-paper) use intent-based activation — they detect the meaning of your request, not specific keywords.
• Claude academic-writing literature-review peer-review ai-research prompt-engineering claude-code academic-pipeline
• Dual-path import infrastructure on all 5 module-level cross-imports (sibling-first, namespace-package fallback) preserves class identity for SemanticScholarUnavailable and bonus-fixes 2 latent-broken import scripts.X paths.

The four option-1..4 trade-offs from the #118 issue body landed on option 2 (new aggregate) — option 4 (re-raise and abort) was rejected for the audit-coverage hit on flaky judge endpoints. Pre-v3.7.3 corpora — run migrate_literature_corpus_to_v3_7_3.py FIRST, then v3.9.0 migration (daisy-chained per spec §3.7; the v3.9.0 tool only acts on entries that already carry contamination_signals.semantic_scholar_unmatched). The policy layer (strict modes, hard-block tier, venue_type / triangulation_policy) is deferred to v3.10 per spec §2.3. Behavioral smoke tests with cross-model spot-check (100% Opus 4.7, ≥75% Sonnet + GPT-5.5). Verifier script scripts/temporal_integrity_audit.py runs 5 passes deterministically. See CHANGELOG.md for the per-issue detail.

Quick Start

V3.5 adds the Collaboration Depth Observer (collaboration_depth_agent, advisory only — never blocks) at every FULL/SLIM checkpoint and at pipeline completion. Stage 2.5 and Stage 4.5 integrity gates run a 7-mode blocking checklist (see academic-pipeline/references/ai_research_failure_modes.md); the reviewer offers an opt-in calibration mode that measures its own FNR/FPR against a user-supplied gold set. V3.8.2 routes those failures through a dedicated uncited_audit_failures aggregate at MED-WARN advisory tier, mirroring the cited path INV-14 row but using a dedicated schema because claim_audit_result.ref_slug is required and the uncited path has no ref to bind. V3.7.3 shipped single-index (Semantic Scholar) contamination detection; v3.9.0 extends to three-index triangulation (S2 + OpenAlex + Crossref) as advisory evidence only.

Architecture & pipeline

Dual-path import infrastructure on all 5 module-level cross-imports (sibling-first, namespace-package fallback) preserves class identity for SemanticScholarUnavailable and bonus-fixes 2 latent-broken import scripts.X paths. Deterministic advisory verifier at the Phase 4 → 5 boundary covering 5 temporal failure modes (P1 retrospective arithmetic, P2 anachronistic citation, P3 comparator unmaterialized, P4 causal inversion, P5 deictic present). F4 (test-count-monotonic harden) reverted because it surfaced pre-existing scripts/ package issue, tracked as #154 (since fixed by PR #158) + re-attempt #155. Socratic mode (deep-research) and Plan mode (academic-paper) use intent-based activation — they detect the meaning of your request, not specific keywords.