Skip to content
TakoAPI
Back to skills

deep-current

by meimakes

Web & Frontend Development
—Not yet rated
Your rating

Brief

Persistent research thread manager with a CLI for tracking topics, notes, sources, and findings.

Description

What This Skill Does

Manages long-running research topics as persistent threads stored in local JSON files. A Python CLI tracks notes, sources, and findings per thread across sessions. Pairs with a nightly cron job to produce dated markdown research reports.

Persistent thread state lets research compound over time rather than restarting from scratch each session.

When to Use It

  • Tracking a fast-moving technology topic across weeks of agent sessions
  • Building a nightly digest on competitors or an emerging market
  • Accumulating sources for a research paper over multiple days
  • Monitoring a developing news story with daily automated updates
  • Maintaining ongoing notes on several parallel research interests

Example Workflow

INPUT

User asks: Add a thread for 'WebAssembly in server-side runtimes' and run a research session

AGENT
  1. 1

    Run python3 scripts/deep-current.py add 'WebAssembly in server-side runtimes' to create the thread

  2. 2

    Check existing threads with list and recent coverage with covered to avoid repeating already-seen sources

  3. 3

    Use web_search and web_fetch to find recent developments on the topic

  4. 4

    Record findings with finding <id> <text> and add sources with source <id> <url>

  5. 5

    Write a dated report to deep-current-reports/YYYY-MM-DD.md with dense findings and inline source links

OUTPUT

A markdown report file in deep-current-reports/ containing findings on the thread, with inline source links and actionable flags

Installation

clawhub install meimakes/deep-current

Run the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.

Links