ai-analyst-lab/ai-analyst-plus
Agent
Agent "google-slides-reviewer" from ai-analyst-lab/ai-analyst-plus, covering agent: google slides reviewer, purpose, inputs, workflow and step 1: read the design spec.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "hypothesis-sharpener" from ai-analyst-lab/ai-analyst-plus, covering agent: hypothesis sharpener, purpose, inputs, output formatting rules and workflow.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "hypothesis" from ai-analyst-lab/ai-analyst-plus, covering agent: hypothesis forming, purpose, inputs, workflow and step 1: parse the question brief.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "narrative-coherence-reviewer" from ai-analyst-lab/ai-analyst-plus, covering agent: narrative coherence reviewer, purpose, inputs, workflow and step 1: headline coherence test.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "notion-export" from ai-analyst-lab/ai-analyst-plus, covering agent: notion export, purpose, inputs, workflow and step 1: read the skill.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "opportunity-sizer" from ai-analyst-lab/ai-analyst-plus, covering agent: opportunity sizer, purpose, inputs, workflow and query logging.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "overtime-trend" from ai-analyst-lab/ai-analyst-plus, covering agent: overtime / trend, purpose, inputs, workflow and pre-flight checks.
ai-analyst-lab/ai-analyst-plus
Agent
Track pipeline execution state for resume capability and progress reporting. Written to working/pipelinestate.json during /run-pipeline execution. Read by /resume-pipeline to determine restart point.
ai-analyst-lab/ai-analyst-plus
Agent
Human-readable summary generated after each pipeline phase completes. Written to working/pipelinesummary.md during execution, updated incrementally. Provides a quick status overview without needing to parse pipelinestate.json.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "presentation-doctor" from ai-analyst-lab/ai-analyst-plus, covering agent: presentation doctor, purpose, inputs, workflow and phase 1: intake (1 min).
ai-analyst-lab/ai-analyst-plus
Agent
Agent "question-framing" from ai-analyst-lab/ai-analyst-plus, covering agent: question framing, purpose, inputs, workflow and step 1: parse and summarize business context.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "receipt-generator" from ai-analyst-lab/ai-analyst-plus, covering agent: receipt generator, purpose, when to run, inputs and workflow.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "root-cause-investigator" from ai-analyst-lab/ai-analyst-plus, covering agent: root cause investigator, purpose, inputs, workflow and pre-flight checks.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "story-architect" from ai-analyst-lab/ai-analyst-plus, covering agent: story architect, purpose, inputs, workflow and phase 1: storyboard (narrative beats).
ai-analyst-lab/ai-analyst-plus
Agent
Agent "story-extractor" from ai-analyst-lab/ai-analyst-plus, covering agent: story extractor, purpose, inputs, workflow and step 1: catalog every data point.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "storytelling" from ai-analyst-lab/ai-analyst-plus, covering agent: storytelling, purpose, inputs, workflow and step 1: ingest the analysis outputs.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "validation" from ai-analyst-lab/ai-analyst-plus, covering agent: validation, purpose, inputs, workflow and step 1: inventory the claims.
ai-analyst-lab/ai-analyst-plus
Agent
Agent "visual-design-critic" from ai-analyst-lab/ai-analyst-plus, covering agent: visual design critic, purpose, inputs, workflow and step 1: load review standards.
Skill Claude CodeCodex
Drive a sealed brief through verdict + readout. Verifies the three-part integrity lock first (R11). SRM check first (R2). Verdict from the decision tree only (R3).
Skill Claude CodeCodex
Replay the decision chain for any experiment in the project. Walks log.md + git log. Default text mode, --diff, or --html.
Skill Claude CodeCodex
Wire a warehouse profile via an interactive wizard. Supports DuckDB, Snowflake, BigQuery, Databricks. Writes /.agentxp/credentials/ / .yaml at chmod 600.
Skill Claude CodeCodex
Drive an experiment from intent through a sealed brief. Pre-registration only — refuses outcome-column queries (R11). Terminates when brief.sealed.yaml lands and the user confirms.
Skill Claude CodeCodex
Walk the renders catalog for an experiment, validate the hash chain, or regenerate the cross-experiment HTML index navigator.
Instructions file
Instructions for ai-analyst-lab/agentxp, covering claude.md — agentxp, 1. identity, 2. skills registry, 3. specialists and 4. the worldview.