Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Agent Research Aggregator

Scans AI agent caches (.claude, .cursor, and more) to turn scattered experiment logs into PaperOrchestra-ready idea.md and experimental_log.md.

Content & WritingAdvanced★ 635⑂ 89AI score 8/10Last updated: Aug 9, 2026

What it does

This skill crawls AI coding-agent cache directories (.claude, .cursor, .antigravity, .openclaw) or any folder you point it at, finds experimentation logs, extracts hypotheses, methods and numeric results, and consolidates them into a single coherent research narrative. The output lands in workspace/inputs/idea.md and workspace/inputs/experimental_log.md, exactly the (I, E) pair the PaperOrchestra paper-writing pipeline expects.

It runs in five phases:

  1. Discovery — discover_logs.py builds a manifest and prints the list of projects found.
  2. Project selection — a paper must come from one project, so the skill always asks you to pick one before any LLM work.
  3. Extraction & synthesis — logs are processed in ~50 KB batches into structured JSON, then merged into one research question, contributions, and result tables.
  4. Formatting — deterministic script writes the two input files with GFM tables.
  5. Audit report — scan statistics, data-quality warnings, and the files still missing.

Safety rules are explicit: never write to cache directories, strip PII and credentials, never fabricate results (low-confidence numbers get an [UNVERIFIED] tag), and confirm with the user when more than 50 files are found.

Who it's for

  • Researchers who ran many agent-driven experiments but kept no structured lab notebook.
  • Grad students racing a conference deadline who need setup sections and result tables fast.
  • Anyone wanting to run PaperOrchestra without existing idea.md / experimental_log.md.
  • Teams whose work history is spread across several agent tools.

Examples

  • "Aggregate my agent logs for paper writing" → the skill scans cwd and ~, shows a numbered project list, and generates both input files for the project you choose.
  • Use --search-roots ~/projects/rl-exp --since 2025-01-01 --depth 3 to limit the scan to one project's recent experiments.
  • If the audit report flags conflicting metrics, inspect the source logs, keep the verified numbers, and re-run the formatting phase.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/Ar9av/PaperOrchestra/HEAD/skills/agent-research-aggregator/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/agent-research-aggregator folder from the GitHub repo Ar9av/PaperOrchestra into my ~/.claude/skills/agent-research-aggregator/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/Ar9av/PaperOrchestra.git && cp -r PaperOrchestra/skills/agent-research-aggregator ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal and move to the folder where you keep tools.
  2. Clone the repository: git clone https://github.com/Ar9av/PaperOrchestra.git
  3. Copy the skill into place: mkdir -p ~/.claude/skills && cp -r PaperOrchestra/skills/agent-research-aggregator ~/.claude/skills/
  4. Verify the scripts/ folder (discover_logs.py, extract_experiments.py, format_po_inputs.py) and references/ prompt files came along; also copy the paper-orchestra skill if you plan to write the full paper.
  5. Check Python 3 is available with python --version.
  6. Restart Claude Code, then ask something like: "Prepare PaperOrchestra inputs from the agent logs in ~/projects/my-experiment" — always name a folder to avoid scanning your whole home directory.
  7. Answer the project-selection question, then review the generated workspace/inputs/idea.md and experimental_log.md.
  8. To continue into paper generation, supply your conference template.tex and conference_guidelines.md.
View source on GitHub ↗License: NOASSERTION