一键导入
diff-runner
Regression testing tool for Oracle of Secrets. Records gameplay scenarios and replays them on new builds to detect logic bugs or desyncs.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Regression testing tool for Oracle of Secrets. Records gameplay scenarios and replays them on new builds to detect logic bugs or desyncs.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use the Agentic File System (AFS) context model for discovery, mounts, scratchpad memory, and file-based tool workflows. Use when a task involves .context operations, AFS CLI usage, or cross-agent context coordination.
Read and interpret the Link to the Past (ALTTP) disassembly with label/symbol context. Use when mapping ROM/CPU addresses to labels, searching usdasm/jpdasm sources, or cross-referencing routines with Hyrule Historian, z3dk label indexes, or z3ed output.
Train, evaluate, and maintain the scawful-echo persona and related avatar models (Echo/Memory/Muse). Use when working on persona voice, dataset prep, A/B testing, deployment, or tool-calling constraints for avatar models.
Autonomous navigation agent for Oracle of Secrets. Uses static map data and dynamic RAM state to localize Link and route him to destinations.
End-to-end Oracle of Secrets debugging using the Mesen2-OOS socket CLI, z3ed, and label/symbol tooling. Use for breakpoint/trace capture, memory inspection, savestate repro, overlays, or crash/regression triage (replaces mesen2-mcp and legacy Lua bridge workflows).
Expert guidance for LLM model training, synthetic data generation (SDE), Hierarchical MoE (H-MoE) management, and agentic evaluation. Use when setting up training runs (Vast.ai), generating datasets, or performing technical research PDF analysis for ROM hacking and general AI development.
| name | diff-runner |
| description | Regression testing tool for Oracle of Secrets. Records gameplay scenarios and replays them on new builds to detect logic bugs or desyncs. |
Execute a scenario script and capture the "Golden Trace".
record scenario.json trace.jsonReplay a trace on the current ROM and assert state matches.
verify trace.json[{"input": "Right", "frames": 60}, ...].record on the stable ROM build.verify on the new ROM build.~/src/hobby/yaze/scripts/ai/diff_runner.py.