
from agent-assistant53
Build and maintain a git-based markdown wiki for teams: ingest, compile, query, and synthesize project knowledge.
LLM Wiki provides a git-backed workflow and CLI-style commands to create, ingest, compile, and query a project knowledge base. It automates setup, deterministic ingestion of source documents, AI-assisted compilation into wiki pages, and query/digest operations with mandatory feedback loops for growing the wiki. The skill encapsulates security rules around untrusted content and safe commit patterns.
Use this skill when a user needs to initialize a project wiki, ingest documents or folders, run compilation to create or update pages, perform targeted queries or deep digests on a topic, lint wiki health, or generate knowledge graphs. It's suitable for teams that track documentation in a repo and want repeatable, auditable wiki updates.
scripts/ for init, ingest, compile, lint, stats, graph, and index updates (has_scripts=true)references/ and example AGENTS.md conventions are provided (has_references=true)/wiki command reference, commit rules, security controls for untrusted content, and operational checks summarized from the SKILL.md bodyAgents capable of running Python-based CLI tooling (Python 3.11+), managing git operations with human confirmation, and performing content synthesis are best suited to this skill.
LLM Wiki is a git-based markdown knowledge base manager for software teams. It provides ingestion (files/URLs to markdown), compilation (AI-driven wiki page generation), querying, linting, and knowledge graph visualization. Scripts are well-structured Python with argparse CLIs, proper error handling, and security-conscious design (prompt injection detection in ingest, untrusted URL flagging). Most scripts require an initialized wiki (.llm-wiki.toml) to run, and init-wiki.py needs a references/ template directory, so they fail gracefully outside a project context. The skill addresses a real and common need — team knowledge management — with a thoughtful workflow.
markitdownpyyamlWell-designed skill with clear security considerations (untrusted content handling, prompt injection detection, user confirmation for git commits). No security concerns. init-wiki.py's TEMPLATES_DIR path resolution would break if scripts are copied without the references/ directory — a packaging/distribution issue rather than a functional bug.
Segment CDP (Analytics.js & Server-side)
Patterns and best-practices for implementing Segment (Analytics.js and server-side) tracking, event schema design, and avoiding common anti-patterns.
Agent Assistant — Debug (Hard)
Deep diagnostic workflow that runs a comprehensive multi-agent debugging routine when a user requests /debug:hard.
Agent Assistant — Business Docs
Workflow to surface and author business documentation when a user requests /docs:business, using the repository's docs commands and core rules.