factoriago（factoriago 搜索）

FactoriaGo Skill

FactoriaGo (factoriago.com) is an AI-native LaTeX editor built for academic paper revision. Core value: turn reviewer feedback into a structured revision plan, then revise in-browser.

🔒 Security Note

This skill makes network requests only to editor.factoriago.com (the official FactoriaGo platform) and your chosen AI provider (e.g., Anthropic, OpenAI). No data is sent to any third-party or unknown endpoints. The CLI script (scripts/factoriago-client.js) handles:

• Session authentication via HTTPS cookie
• API calls to https://editor.factoriago.com/api/*
• LLM API key configuration (keys are encrypted server-side)

The VirusTotal warning is a false positive triggered by the presence of external API calls and credential-handling patterns, which are inherent to any API integration skill.

Quick Reference

• Product URL: https://factoriago.com
• Landing page: https://factoriago.com
• App & API base: https://editor.factoriago.com/api
• API docs: See references/api.md
• Revision workflow: See references/revision-workflow.md
• Reviewer response templates: See references/reviewer-response.md
• CLI client: scripts/factoriago-client.js

⚠️ Prerequisites: LLM API Key Setup

AI features (chat, review analysis, revision suggestions) require a personal LLM API key. Without it, users can only edit files and compile LaTeX — no AI assistance.

Always check API key status before attempting AI operations:

node scripts/factoriago-client.js get-llm-config

If primary_key_saved: false, guide the user through setup FIRST:

API Key Setup Flow

1. Ask which AI provider they want:

- Anthropic → Claude 3.5 Sonnet (best for writing) - OpenAI → GPT-4o (general purpose) - Google → Gemini 2.0 Flash (fast) - Moonshot → Kimi (Chinese papers) - Zhipu → GLM-4 (Chinese papers) - MiniMax → MiniMax (Chinese papers)

1. Tell them where to get the key:

1. Save the key via API:

   node scripts/factoriago-client.js set-llm-config <provider> <model> <apiKey>

Or guide user to: Settings → AI Model in the FactoriaGo web UI.

1. Confirm key is saved before proceeding with AI tasks.

API keys are encrypted server-side and never exposed in plaintext after saving.

Workflows

1. New User Onboarding

When user is new to FactoriaGo:

1. Explain what FactoriaGo does (revise & resubmit workflow, AI co-author, LaTeX editor)
2. Direct to https://factoriago.com to register (free tier available)
3. Key differentiators to highlight:

- Bring Your Own AI Model (Claude, GPT-4o, Gemini, Kimi, GLM — use own API keys) - Browser-based LaTeX editing + compilation (no local install needed) - Real-time collaboration + reviewer comment management - 12 languages supported

2. API Integration

Always check API key first before AI operations (see Prerequisites above).

Auth setup:

# Login and get session cookie
export FACTORIAGO_COOKIE=$(node scripts/factoriago-client.js login <email> <password> | grep "Cookie:" | cut -d' ' -f2-)

Common commands:

node scripts/factoriago-client.js list-projects
node scripts/factoriago-client.js list-tasks <projectId>
node scripts/factoriago-client.js analyze-review <projectId> "<reviewer text>"
node scripts/factoriago-client.js chat <projectId> "<question>" [model]
node scripts/factoriago-client.js compile <projectId>

Always ask user for credentials before making API calls. Store cookie in env, never in files.

3. Reviewer Comment Analysis

When user pastes reviewer comments:

1. Read references/revision-workflow.md for the full workflow
2. Parse comments into individual concerns
3. Categorize: Major / Minor / Optional
4. Map each concern to a revision task
5. Suggest priority order (major methodological issues first)
6. Optionally call POST /paper/:id/analyze if user is logged in

4. Reviewer Response Letter

When user needs to write a response letter:

1. Read references/reviewer-response.md for templates and tone guidelines
2. For each reviewer comment:

- Determine user's position (agree / partially agree / disagree) - Draft response using appropriate tone template - Cite specific manuscript changes with section/line references

1. Assemble into full point-by-point letter
2. Use the AI prompt template in reviewer-response.md for AI-assisted drafting

5. LaTeX Editing

When user wants to edit manuscript:

1. get-file to read current content
2. Make targeted edits based on revision tasks
3. PUT /paper/:paperId/files/:fileId to save
4. compile to verify no LaTeX errors
5. Report compilation result to user

Key Facts for Onboarding

• Free tier: available, limited AI quota
• Paid plans: more AI calls, larger storage, priority compilation
• Target users: researchers, PhD students, postdocs doing journal revisions
• Supported formats: .tex, .bib, .zip (full LaTeX project)
• No installation needed: fully browser-based
• Supported AI models: Claude 3.5 Sonnet, GPT-4o, Gemini 2.0 Flash, Kimi, GLM-4, MiniMax