skills.rest
shenmuxing
shenmuxingCOMMUNITY
Explore all skills by this author →
GitHub

deepseek-agent

Delegate critique, revision, and manuscript-writing tasks to a DeepSeek-backed OpenCode agent.

2|1|Updated May 7, 2026
Search Tags:#agent delegation#deepseek#opencode#academic writing#manuscript revision#research critique
One-click install
npx skills add https://github.com/shenmuxing/EtaSkill --skill deepseek-agent
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deepseek-agent
Source: https://github.com/shenmuxing/EtaSkill/tree/main/skill-examples/deepseek-agent
Command: npx skills add https://github.com/shenmuxing/EtaSkill --skill deepseek-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill helps Codex hand off demanding critique, revision, and manuscript-writing work to a separate DeepSeek-backed agent while keeping Codex in control of the workflow and final review.

Core Features & Use Cases

  • External agent delegation: Prepare a brief, invoke OpenCode, and collect DeepSeek output as a separate agent pass.
  • Writing and revision support: Useful for adversarial research critique, novelty skepticism, method review, paper prose, LaTeX section drafting, and iterative academic text revision.
  • Operational guardrails: Includes preflight checks, session reuse guidance, result-file handling, timeout controls, and verification steps for safer non-interactive use.
  • Example use case: Use this Skill when you want a second-pass reviewer to challenge a draft, rewrite a section, or produce structured feedback before you integrate changes.

Quick Start

Ask Codex to delegate your critique or revision task through the deepseek-agent skill, providing the target files, constraints, and desired output contract in a brief.

Frequently Asked Questions about deepseek-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I delegate academic manuscript writing and critique to DeepSeek?▼

To delegate academic manuscript revision to DeepSeek, you ask Codex to invoke the deepseek-agent skill with your target files, constraints, and desired output contract, which OpenCode then processes and returns.

What is adversarial research review and how does an agent perform novelty checking?▼

Adversarial research review challenges a draft's novelty and methodology using an external agent. The DeepSeek-backed agent acts as a second-pass reviewer to provide structured critique and skepticism before you integrate changes.

Do I need OpenCode and PowerShell to run DeepSeek manuscript revision tasks?▼

Yes, you need OpenCode and PowerShell configured with valid DeepSeek credentials to run manuscript revision tasks. The skill relies on these dependencies for preflight checks, timeout controls, and non-interactively handling result files.

Can I reuse a session for iterative academic text revision and LaTeX drafting?▼

Yes, you can reuse sessions for iterative academic text revision and LaTeX section writing. The skill includes session reuse guidance and wrapper support for transcripts, allowing continuous critique and drafting loops.

What are the limitations of using a DeepSeek-backed agent for research critique?▼

Limitations include dependency on configured DeepSeek credentials and OpenCode availability. The skill requires preflight checks and timeout controls to manage non-interactive use safely, meaning environment misconfigurations will halt execution.

Related Skills

Explore More Education & Research Skills→
chiptoe-svgchiptoe-svg

add-opencode

Integrate the OpenCode agent provider into NanoClaw for alternative LLM backends.

Community
Intermediate
Alexander-M-DickersonAlexander-M-Dickerson
58

edit-section

Rewrite LaTeX sections to improve style and citation consistency.

Community
Advanced
cyberAlchemyAIcyberAlchemyAI
20

invoke-example-runner

Execute invoke template validation prompts in Codex and save Invoke Results.

Official
Advanced
Iron-HamIron-Ham
31

deep-review

Run parallel multi-agent code quality reviews and output prioritized reports.

Community
Advanced