Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

research-mode

Investigation workflow — gather sources, cross-check, summarize, cite. Use for literature review, tech comparisons, CVE research, and "how does X work" questions that need evidence.

Exécuter dans Manus

Étoiles1

Forks0

Mis à jour24 avril 2026 à 16:02

Source

jmoraleses

jmoraleses/Ey-code

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Commande d'installation

Téléchargement

Exécuter dans Manus

Utile pourSOC

Développeurs de logicielsProfessions informatiques et mathématiques15-1252L4

SKILL.md

readonly

name	research-mode
description	Investigation workflow — gather sources, cross-check, summarize, cite. Use for literature review, tech comparisons, CVE research, and "how does X work" questions that need evidence.
when_to_use	- User asks "research X" or "compare A vs B". - User asks for CVEs, vulnerabilities, or security advisories on a component. - User asks to summarize a paper, repo, or long document.
license	MIT

Research Mode

Pipeline

Frame — restate the question in one sentence. List 3 sub-questions.
Gather — web_search, read_file on local refs, github_info for repos.
Cross-check — require ≥2 independent sources for any non-trivial claim.
Synthesize — write findings as: claim → evidence → source.
Cite — every factual claim has a URL or file:line. No citation, no claim.
Deliver — markdown report to .reports/<topic>.md if the work is non-trivial.

Rules

Distinguish what the sources say from your interpretation. Mark interpretation with "Reading:".
Admit unknowns. If sources conflict, say so and stop. Don't invent a tiebreaker.
Date-stamp everything. Security findings and benchmarks go stale fast.
No hallucinated citations. If you can't produce a URL, don't quote it.

Output template

# <topic>

## Question
<one sentence>

## Findings
- <claim> — [source](url)
- <claim> — [source](url)

## Reading (interpretation)
<what this means for the user's actual need>

## Open questions
- <what's still unresolved>

Preferred tools

web_search, github_info, read_file, semantic_search, search_references, add_reference.

Plus depuis ce dépôt

même dépôt

ai-math-research

jmoraleses/Ey-code

Conduct advanced research in Artificial Intelligence and Mathematics, including the ability to generate and run automated tests or proofs.

2026-04-241

auto-test

jmoraleses/Ey-code

Automatic test generation and execution. Use whenever you need to verify code functionality, ensure quality, or validate implementations through comprehensive testing.

2026-04-241

coding-agent

jmoraleses/Ey-code

Programming workflow for the runtime — read before edit, surgical changes, test-driven fixes, and efficient use of existing tools.

2026-04-241

ey-agent-core

jmoraleses/Ey-code

Core behavioral contract for the Ey-Code agent runtime. Use always — defines how the model coordinates reasoning and tool usage.

2026-04-241

karpathy-guidelines

jmoraleses/Ey-code

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

2026-04-241

pentesting-chat

jmoraleses/Ey-code

Ethical pentesting operations directly from natural language chat, including the ability to install necessary security tools in a secure environment.

2026-04-241

name	research-mode
description	Investigation workflow — gather sources, cross-check, summarize, cite. Use for literature review, tech comparisons, CVE research, and "how does X work" questions that need evidence.
when_to_use	- User asks "research X" or "compare A vs B". - User asks for CVEs, vulnerabilities, or security advisories on a component. - User asks to summarize a paper, repo, or long document.
license	MIT

Research Mode

Pipeline

Frame — restate the question in one sentence. List 3 sub-questions.
Gather — web_search, read_file on local refs, github_info for repos.
Cross-check — require ≥2 independent sources for any non-trivial claim.
Synthesize — write findings as: claim → evidence → source.
Cite — every factual claim has a URL or file:line. No citation, no claim.
Deliver — markdown report to .reports/<topic>.md if the work is non-trivial.

Rules

Distinguish what the sources say from your interpretation. Mark interpretation with "Reading:".
Admit unknowns. If sources conflict, say so and stop. Don't invent a tiebreaker.
Date-stamp everything. Security findings and benchmarks go stale fast.
No hallucinated citations. If you can't produce a URL, don't quote it.

Output template

# <topic>

## Question
<one sentence>

## Findings
- <claim> — [source](url)
- <claim> — [source](url)

## Reading (interpretation)
<what this means for the user's actual need>

## Open questions
- <what's still unresolved>

Preferred tools

web_search, github_info, read_file, semantic_search, search_references, add_reference.