dify-docs-guides
A structured workflow for maintaining and writing Dify technical documentation.
Install
mkdir -p .claude/skills/dify-docs-guides && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14685" && unzip -o skill.zip -d .claude/skills/dify-docs-guides && rm skill.zipInstalls to .claude/skills/dify-docs-guides
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Use when writing, improving, or reviewing Dify user guide documentation. Covers pages in en/{cloud,self-host}/use-dify/, en/develop-plugin/, and en/self-host/deploy/. Triggers: "write docs for [feature]", "improve this page", "review this documentation section".Key capabilities
- →Read Dify documentation style, formatting, and glossary guides
- →Research and verify feature behavior against the Dify codebase
- →Identify target pages and match them to reader personas
- →Draft documentation content for specific personas
- →Translate English changes to Chinese and Japanese
How it works
The skill starts by reading style, formatting, and glossary guides. It then researches and verifies feature behavior against the Dify codebase, drafts content for specific personas, and translates changes into Chinese and Japanese. Finally, it runs a check chain for format, terminology, and reader tests.
Inputs & outputs
When to use dify-docs-guides
- →Write docs for a new feature
- →Improve an existing documentation page
- →Review a documentation section
About this skill
User Guide Rules
Not an entry point — run under dify-docs-write; this pack's rules and procedures implement its stages for user-guide pages.
Scoping rules (S1)
Match each target page to its reader persona (see Reader Personas below). use-dify pages exist as two product copies — en/cloud/use-dify/ and en/self-host/use-dify/ — with no shared pages and no cross-audience navigation. When the page exists in both, scope both: shared-content improvements land in both copies in the same pass; audience-specific blocks (plan gating, env-var callouts, Enterprise tips) stay per-copy.
S2 discovery — environment variables
Check for feature-related environment variables. In the Dify codebase, run:
grep -rn "<FEATURE_KEYWORD>" docker/.env.example docker/envs/ api/configs/
Use the feature's name as it would appear in a variable (e.g. COLLABORATION); try 2–3 keyword variants before concluding. No matches → the feature has no env-var surface; skip the env-var guidance below. Matches → record each variable as mandatory or optional plus its default, and queue the dify-docs-env-vars pack's procedure in the S4 scope report, to update en/self-host/deploy/configuration/environments.mdx in the same session — that reference is the single source of truth for variable semantics.
Drafting rules (S5)
- Write for the persona of the page's path (see Reader Personas below).
- Never restate the page's own audience. Everyone on a self-host page is self-hosted and everyone on a cloud page is on Dify Cloud, so "On self-hosted deployments, …" and "On Dify Cloud, …" are banned on their own pages. Audience qualifiers are legitimate only on the audience-neutral trees (
en/learn/,en/api-reference/,en/cli/,en/develop-plugin/), where they genuinely disambiguate. Two carve-outs are deliberate and stay: naming a different product (the<Tip>surfacing Dify Enterprise where a CE capability ends — see "Paid Feature Callouts" in the style guide), and a comparative-advantage claim, where the qualifier marks something this product has that the other lacks ("On Dify Cloud, many popular trigger integrations are pre-configured" — the point is the perk, not the scope). Scoping restatement is banned; advantage framing is not. - If S2 discovery found related environment variables, present them per Environment Variables in User Guides below.
Reader Personas
Adjust tone and assumed knowledge by document path. This is the master copy of these personas; other skills point here.
| Path | Readers | Assume | Prioritize |
|---|---|---|---|
en/cloud/use-dify/ | Dify Cloud app builders; mix of developers and non-technical users | Basic AI familiarity; no infrastructure or deep coding knowledge — there is no deployment to operate | Task completion and outcomes; explain technical concepts when they appear |
en/self-host/use-dify/ | App builders whose team self-hosts Dify | Basic AI familiarity; no deep coding knowledge. Their deployment exists, so env vars are fine to mention — but the reader is a builder, not the operator | Task completion and outcomes; explain technical concepts when they appear; keep deployment mechanics one click away in the env reference |
en/self-host/deploy/ | DevOps engineers and system administrators | Strong infrastructure knowledge: Docker, databases, networking, environment variables | Precise technical detail; don't over-explain standard operations |
en/develop-plugin/ | Developers building custom Dify plugins | Strong Python skills; familiarity with Dify's core concepts | API contracts, extension points, code patterns; code examples are essential |
Collaboration Model
The user brings documentation expertise and user empathy; you bring AI domain knowledge. Apply it actively:
- Explain why an AI concept is designed the way it is and what problem it solves, not just what it does (e.g., explain a tool role in conversation history at the LLM API mechanism level).
- When the user questions a product design, say whether it is Dify-specific or an industry norm, and how users are likely to understand it.
- Replace abstract explanations with concrete scenarios that show why a user needs the feature.
- When the user is unsure about phrasing, judge it from the reader's side: is the term understandable, accurate in the AI context, and close to the user's mental model?
- Flag unusual designs, likely misunderstood concepts, and inaccurate AI-domain terms unprompted.
Environment Variables in User Guides
Applies when S2 discovery found related variables. The two product copies fork — never mix the patterns:
- Self-host copy (
en/self-host/use-dify/): name the mandatory variables and the values to set in a callout (rules below), and link to the reference.environments.mdx(maintained via thedify-docs-env-varspack) owns everything else: defaults, mechanisms (worker classes, proxy paths, scheme rules, fallbacks), interactions, and failure modes. - Cloud copy (
en/cloud/use-dify/): never any env-var content. A feature that is simply on in Dify Cloud gets nothing; a plan-limited feature gets the plan-gating pattern (<Badge color="blue">Professional</Badge> and <Badge color="blue">Team</Badge>in prose with a Learn more link — see "Paid Feature Callouts" in the style guide).
In the self-host copy:
-
Place the configuration in a callout, never a dedicated H2 section. Configuration enablement is an aside to the page's task flow, and the defaults and mechanics live one click away in the env reference.
-
Pick the callout type:
<Note>when the variables are mandatory (the feature does not work at all without them);<Info>when they only customize behavior that already works. -
Use this pattern — open with the feature state, never an audience qualifier (Drafting rule 2):
<Note> [Feature] is off by default. Enable it by setting: - `VAR_NAME` = `value` See [Environment Variables](/en/self-host/deploy/configuration/environments#var_name) for details. </Note> -
Exclude from the user guide: default values, the "why" behind each variable, deployment-specific mechanics, and variable interactions — anything the reader finds one click into the reference.
When not to use it
- →When the task is not related to Dify user guide documentation
- →When the user wants to document features without verifying behavior against code
- →When the user is asking to write documentation for a different product
Limitations
- →The skill is specific to Dify user guide documentation.
- →It requires access to the Dify codebase for verification.
- →The skill mandates translation into Chinese and Japanese for every English change.
How it compares
This skill enforces a strict, multi-step process for Dify documentation, including mandatory research against the codebase, persona-based writing, and multi-language translation, providing a highly structured and verifiable approach compare
Compared to similar skills
dify-docs-guides side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| dify-docs-guides (this skill) | 0 | 1mo | Review | Advanced |
| documentation | 0 | 4mo | Review | Beginner |
| prd | 0 | 4mo | No flags | Intermediate |
| postmortem-writing | 17 | 2mo | No flags | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by langgenius
View all by langgenius →You might also like
documentation
popo0407
ドキュメント作成とナレッジ管理の原則
prd
alfredolopez80
Product Requirements Document generation and management with INVEST-compliant user stories
postmortem-writing
wshobson
Write effective blameless postmortems with root cause analysis, timelines, and action items. Use when conducting incident reviews, writing postmortem documents, or improving incident response processes.
internal-comms
anthropics
A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).
wiki-page-writer
microsoft
Generates rich technical documentation pages with dark-mode Mermaid diagrams, source code citations, and first-principles depth. Use when writing documentation, generating wiki pages, creating technical deep-dives, or documenting specific components or systems.
confluence-assistant
tech-leads-club
Expert in Confluence operations using Atlassian MCP - automatically detects workspace Confluence configuration or prompts for site details. Use for searching, creating, updating pages, managing spaces, and adding comments with proper Markdown formatting.