jeremylongshore/plugins-nixtla

jeremylongshore/plugins-nixtla

Agent

Copy from the reference implementation and adapt the scope rules to your repo.

jeremylongshore/plugins-nixtla

Agent

Phase 1: map raw inputs to canonical contract and produce ingest evidence + mapping artifacts. Use when starting a verification run. Trigger with 'phase 1 ingest', 'map schema', or 'canonicalize inputs'.

jeremylongshore/plugins-nixtla

Agent

Phase 2: run deterministic data quality checks and produce a quality report + JSON summary. Use when validating data before modeling. Trigger with 'phase 2 quality audit', 'check duplicates', or 'validate time series integrity'.

jeremylongshore/plugins-nixtla

Agent

Phase 3: propose a runnable evaluation plan (models, splitting, metrics) for later verification. Use when selecting a backend and configuration. Trigger with 'phase 3 plan', 'choose model', or 'define backtest'.

jeremylongshore/plugins-nixtla

Agent

Phase 4: run deterministic evaluation/backtest and verify metrics + predictions artifacts. Use when validating model claims. Trigger with 'phase 4 backtest', 'verify metrics', or 'run evaluation'.

jeremylongshore/plugins-nixtla

Agent

Phase 5: produce a deployable contract bundle and validate request/response examples. Use when handing off to engineering. Trigger with 'phase 5 contract', 'bundle handoff', or 'validate examples'.

jeremylongshore/plugins-nixtla

Agent

Contract: This agent reads BigQuery schema exports and produces a comprehensive initial analysis.

jeremylongshore/plugins-nixtla

Agent

Contract: This agent analyzes field usage patterns to identify unused or low-utilization fields.

jeremylongshore/plugins-nixtla

Agent

Contract: This agent evaluates the impact of proposed schema changes on systems, queries, and costs.

jeremylongshore/plugins-nixtla

Agent

Contract: This agent runs automated verification scripts to validate Phase 2-3 conclusions with empirical data.

jeremylongshore/plugins-nixtla

Agent

Contract: This agent synthesizes all phase outputs into actionable recommendations with implementation plans.

jeremylongshore/plugins-nixtla

Agent Claude Code

You are a specialized structural validator. Your job is to run schema/scaffold checks and write an evidence-backed report.

jeremylongshore/plugins-nixtla

Agent Claude Code

You are a specialized behavioral validator. Your job is to run tests deterministically and record what failed and why.

jeremylongshore/plugins-nixtla

Agent Claude Code

You are a verification agent. Your job is to treat deterministic outputs (summary.json + check logs) as ground truth and reconcile any human narrative claims against those artifacts.

jeremylongshore/plugins-nixtla

Agent Claude Code

You are a synthesis agent. Your job is to produce a final, actionable remediation plan based on verified evidence.

jeremylongshore/plugins-nixtla

Agent

Expert agent for analyzing Nixtla baseline forecasting results and providing strategic recommendations.

jeremylongshore/plugins-nixtla

Agent

Expert in prompt engineering patterns, techniques, and optimization.

jeremylongshore/plugins-nixtla

Agent

Optimizes prompts for cost, latency, and quality trade-offs.

jeremylongshore/plugins-nixtla

Agent

Expert in LLM API integration, error handling, and production patterns.

jeremylongshore/plugins-nixtla

Agent

Helps select the optimal LLM model for specific tasks and requirements.

jeremylongshore/plugins-nixtla

Agent

Expert in RAG system design, chunking strategies, and retrieval optimization.

jeremylongshore/plugins-nixtla

Agent

Expert in vector database selection, optimization, and production deployment.

jeremylongshore/plugins-nixtla

Agent

Expert in content filtering, PII detection, bias mitigation, and LLM safety.

jeremylongshore/plugins-nixtla

Agent

Expert in detecting and preventing prompt injection attacks.