Skill Claude CodeCodex
Codex-native adapter for invoking Anthropic Claude Code CLI as an independent read-only QAQC reviewer using markdown file handoff. Use when the user explicitly asks Codex to ask Claude, Claude Code, Opus, or Anthropic for a QAQC/code review, second-model review, or independent review. Do not trigger for generic QAQC…
Agent Claude Code
This directory contains definitions for specialist agents that handle specific domains within ras-commander workflows.
Agent Claude Code
Enforces ras-commander API consistency patterns through static analysis guidance. Detects violations: missing @staticmethod, @logcall, incorrect parameter naming, inflexible path typing, classes with init that should be static. Use for: API review, pattern enforcement, violation detection, fix suggestions. Triggers…
Agent Claude Code
Extracts best practices and successful strategies from conversation history. Use to identify patterns worth documenting and formalizing. Triggers: "find best practices", "successful patterns", "what works", "extract lessons".
Agent Claude Code
Detects recurring blockers and their resolutions in conversation history. Use to identify common issues and prevention strategies. Triggers: "find blockers", "identify issues", "recurring problems", "troubleshooting patterns".
Agent Claude Code
Expert in Claude Code best practices, configuration, and official Anthropic documentation. Consults official docs for skills creation, memory hierarchy (CLAUDE.md, .claude/rules/), imports, path-specific rules, and Claude Code configuration. Use when implementing Claude Code features, creating skills, organizing…
Agent Claude Code
Deep code planning and review oracle using OpenAI Codex CLI (gpt-5.2-codex). Supports TWO invocation patterns: 1. Markdown file handoff (recommended) - Write TASK.md, execute, read OUTPUT.md 2. Direct CLI with HEREDOC (quick tasks) - codex e "prompt" --full-auto Best for tasks requiring 20-30 minutes of deep analysis…
Agent Claude Code
Legacy Claude-only code analysis oracle using Google Gemini CLI. This is not part of the repository's standard production harness policy, which is Claude Code plus Codex. Keep only for historical Claude-only workflows where the user explicitly requests Gemini. Supports TWO invocation patterns: 1. Markdown file handoff…
Agent Claude Code
Expert-level deep analysis of conversation history for strategic insights. Use for complex synthesis, cross-conversation analysis, and strategic recommendations. Triggers: "deep analysis", "strategic insights", "expert review", "synthesize learnings".
Agent Claude Code
Fast scanner for Claude Code conversation history index. Use for quick analysis of history.jsonl to extract prompts, projects, and activity patterns. Triggers: "scan history", "list conversations", "prompt frequency", "project activity".
Agent Claude Code
Orchestrates conversation history analysis by coordinating specialized sub-agents. Use when analyzing Claude Code conversation history for insights, patterns, and recommendations. Triggers: "analyze conversations", "conversation insights", "learn from history", "extract patterns", "find slash commands", "identify…
Agent Claude Code
Expert librarian for ras-commander’s example notebooks in examples/.ipynb. Maintains notebook conventions, helps users author new notebooks using proven patterns, and autonomously QA/QC’s the example suite as a functional test harness (run → review → fix → document). Uses notebook-runner to execute notebooks and…
Agent Claude Code
Retrieves and summarizes official HEC-HMS documentation from www.hec.usace.army.mil to ground-truth HMS↔RAS linked workflows, Atlas/TP-40 discussions, and user-facing notebook guidance. Produces short, link-rich “ground truth” notes with key quotes and version context. Use when users mention: HEC-HMS manual, HMS…
Agent Claude Code
HEC-RAS code archaeology agent for reverse engineering .NET assemblies, understanding internal algorithms, and documenting undocumented behavior. Triggers: "analyze assembly", "decompile", "code archaeology", "internal algorithm", "understand HEC-RAS [feature]", "find algorithm for", "investigate internals", "discover…
Agent Claude Code
Thin coordinator for HEC-RAS project workflows. Dispatches to specialist agents and skills, then aggregates results into unified reports. Main entry point for complete HEC-RAS project operations: inspect, plan, execute, analyze. Use when running full project workflows, executing all plans, performing end-to-end…
Agent Claude Code
Notebook-driven HEC-RAS project QA/QC reviewer. Reads notebook code cells to infer expected HEC-RAS project folders/files, then verifies project linkages, file targeting, unsteady boundary locations, plan→unsteady bindings, and HDF comparison methodology soundness. Use when: notebook QAQC, boundary-condition…
Agent Claude Code
Loads and analyzes HEC-RAS projects to produce actionable intelligence reports. Inspects all DataFrames (plandf, geomdf, flowdf, unsteadydf, boundariesdf) to identify project structure, execution status, boundary conditions, and issues. Use when initializing projects, auditing project state, discovering runnable…
Agent Claude Code
Interprets HEC-RAS simulation results beyond raw data extraction. Verifies execution success via compute message parsing, extracts key hydraulic metrics (max WSE, velocities, peak timing), compares against expected thresholds, and flags anomalies (unrealistic values, missing data, instabilities). Use when validating…
Agent Claude Code
Curates the repository's AGENTS-first instruction graph, Claude-native agents, skills, rules, generated Codex skill bridge, and agenttasks memory system. Use when organizing project memory, creating or auditing skills and agents, consolidating task outputs, refactoring instruction files, or checking that Claude Code…
Agent Claude Code
Fast reviewer for condensed notebook output digests (audit.md / audit.json). Flags “unexpected behavior” signals even when no exception is raised: empty results, missing expected artifacts, suspicious NaNs/zeros, path leakage, and warnings that indicate degraded correctness. Use as a downstream reviewer for…
Agent Claude Code
Fast reviewer for condensed notebook output digests (audit.md / audit.json). Finds exceptions, tracebacks, stderr, and failing cells; reports exact cell indices and likely root causes. Use as a downstream reviewer for notebook-runner and example-notebook-librarian when notebooks are large.
Agent Claude Code
Runs and troubleshoots Jupyter notebooks (.ipynb) as repeatable tests and executable documentation. Uses papermill as primary execution engine for parameterization, progress reporting, and robust error handling. Falls back to nbconvert or pytest+nbmake for specific scenarios. Use when users say: run notebook, execute…
Agent Claude Code
Automated setup and management of Python environments for ras-commander. Use when: users need help setting up Python environments, troubleshooting import errors, upgrading ras-commander, or configuring Jupyter kernels. Trigger phrases: "set up environment", "install ras-commander", "conda create", "import error"…
Agent Claude Code
Expert in ras-commander API, dataframe structures, and integration patterns. Spawns explore subagents to search the large/sprawling library codebase. Knows RasPrj dataframe columns (plandf, geomdf, flowdf, etc.) and return formats from all functions. Use when integrating with ras-commander, understanding API patterns…