跳转到主要内容
此内容尚未提供您的语言版本,正在以英文显示。

Ara Compiler

技能 已验证 活跃

Compiles any research input — PDF papers, GitHub repositories, experiment logs, code directories, or raw notes — into a complete Agent-Native Research Artifact (ARA) with cognitive layer (claims, concepts, heuristics), physical layer (configs, code stubs), exploration graph, and grounded evidence. Use when ingesting a paper or codebase into a structured, machine-executable knowledge package, building an ARA from scratch, or converting research outputs into a falsifiable, agent-traversable form.

目的

To transform raw research materials into a falsifiable, agent-traversable knowledge package, enabling structured analysis and reproducibility.

功能

  • Compiles diverse research inputs (PDFs, repos, logs, notes)
  • Generates ARA with cognitive and physical layers
  • Creates exploration graph and grounded evidence
  • Handles partial inputs gracefully and asks for clarification
  • Validates generated ARA artifacts

使用场景

  • Ingesting a research paper or codebase into a structured knowledge package.
  • Building an Agent-Native Research Artifact from scratch.
  • Converting research outputs into a falsifiable, agent-traversable form.
  • Organizing and structuring scattered research notes and findings.

非目标

  • Generating novel research claims or hypotheses not present in the source material.
  • Performing active research or experimentation beyond artifact compilation.
  • Replacing the human researcher's critical analysis or interpretation.

工作流

  1. Read all provided inputs thoroughly.
  2. Reason through the 4-stage epistemic chain-of-thought (deconstruction, mapping, stubbing, graph extraction).
  3. Generate all mandatory ARA files.
  4. Perform a coverage check loop (max 3 rounds) to identify and patch gaps.
  5. Validate the ARA using Seal Level 1 checks.
  6. Fix any validation failures and re-validate.
  7. Report a summary of the generated artifact and validation result.

Practical Utility

  • info:Usage examplesWhile the SKILL.md describes input handling strategies, concrete, end-to-end runnable examples demonstrating invocation and observable outcomes are not explicitly provided.

安装

请先添加 Marketplace

/plugin marketplace add Orchestra-Research/AI-Research-SKILLs
/plugin install AI-Research-SKILLs@ai-research-skills

质量评分

已验证
95 /100
1 day ago 分析

信任信号

最近提交17 days ago
星标8.3k
许可证MIT
状态
查看源代码