Index of every page on this site: llms.txt
All skills
sanyuan0704 avatar

/wiki-ingest

@08b6572
Run it yourself

Compile articles, documents, or notes into a structured wiki knowledge base. Use when user says 'ingest to wiki', 'compile to knowledge base', 'update wiki', 'wiki ingest', 'add this to wiki', or invokes /wiki-ingest. Supports single or batch ingest. Triggers: wiki, ingest, knowledge base, compile, digest, index, catalog.

Use this Skill: https://skilld.dev/gh/sanyuan0704/sanyuan-skills/wiki-ingest

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈84 tokens for metadata: the name and description. ≈893 when used: this file. ≈369 more on demand in 1 file.

Description uses 4.2% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈337 characters, or 4.2%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Wiki Ingest — Knowledge Base Compiler

Compile any text content (articles, documents, notes) into structured wiki pages with cross-references, building a searchable, interconnected knowledge network.

IRON LAW: One wiki page = one knowledge entity. Never cram multiple concepts into one page.

Wiki Directory Structure

Wiki defaults to wiki/ under the current project root. User may specify a different path.

wiki/
├── index.md              # Full table of contents
├── log.md                # Append-only operation log
├── concepts/             # Core concept pages
├── products/             # Product/tool entity pages
├── patterns/             # Engineering patterns & design decisions
└── comparisons/          # Cross-topic comparison pages

Workflow

1. Confirm Input

Accept user-specified content source:

  • Pasted text directly
  • File path(s) (md, txt, pdf, etc.)
  • Directory path (batch processing)

If user doesn't specify a wiki path, use wiki/. Create the directory if it doesn't exist.

2. Check Existing Wiki

Read wiki/index.md (if exists) to understand existing pages and avoid duplicates.

3. Extract Knowledge Entities

Extract from content:

  • Concepts (concepts/): Abstract ideas, terminology, theories
  • Products (products/): Specific products, tools, services
  • Patterns (patterns/): Engineering patterns, design decisions, methodologies
  • Comparisons (comparisons/): Cross-product or cross-approach analysis

Extraction threshold: An entity deserves its own page only if it would be referenced by other pages.

4. Create or Update Pages

Each entity maps to one wiki page.

→ Load references/page-templates.md for page templates.

Rules:

  • If page already exists → update it, append new information, don't overwrite existing content
  • Keep the one-line definition stable unless the new content provides a clearly better one
  • Prioritize updating "Sources" and "Related Pages" sections

5. Update Cross-References

Check all existing wiki pages. If new content involves concepts referenced in other pages:

  • Update their Related Pages list
  • Update their Sources list
  • Supplement new information in their detail sections

6. Update index.md

Add new page entries to wiki/index.md, organized by category.

7. Append to log.md

Append to wiki/log.md:

## YYYY-MM-DD: Ingest <source title>

**Source**: <file path or "user input">
**New pages**: list newly created pages
**Updated pages**: list modified pages
**New cross-references**: list newly established links

8. Report Results

Tell user: what pages were created, what pages were updated, what new cross-references were established.

File Naming

  • All lowercase, hyphen-separated: agent-loop.md, kv-cache.md
  • Name by entity, not by number

Cross-Reference Format

  • Use [[category/page-name]] double-link format (Obsidian-compatible), e.g. [[concepts/agent-loop]], [[products/claude-code]]
  • Always include the category prefix to avoid ambiguity across directories
  • List all related pages at the bottom of each page

Guidelines

  • Never create duplicate pages — always check wiki/ first
  • Don't extract trivial entities — if a concept appears once and won't be referenced elsewhere, skip it
  • For batch processing: extract all entities first, then create pages and update references in one pass to avoid inconsistent intermediate states
  • Always update index.md and log.md after every ingest

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts2mo4 checks · Risk SAFE
  • Gen Agent Trust Hub2mo

    The skill is designed to compile and organize documents into a wiki knowledge base. While generally safe, it possesses a vulnerability to indirect prompt injection because it processes untrusted user-provided content without using boundary markers or sanitization, which could allow instructions hidden in those documents to influence the agent's behavior.

  • Socket2mo

    No alerts

  • Snyk2mo

    Risk: LOW · No issues

  • ZeroLeaks7mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 08b6572. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 months ago.

Dormantupdated 7 months ago

README badge

README badge for sanyuan0704/sanyuan-skills/wiki-ingest