geronimo-iia/llm-wiki-skills/config

Claude Code and agent skill by geronimo-iia/llm-wiki-skills

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Read, write, and understand wiki configuration — global defaults, per-wiki overrides, and resolution order.

Install

npx vskill@latest install geronimo-iia/llm-wiki-skills/config

SKILL.md

Config

Two configuration files control llm-wiki behavior:

  • ~/.llm-wiki/config.toml — global engine config (local to the

machine, never committed). Override the path with --config <path> or the LLM_WIKI_CONFIG environment variable.

  • <wiki>/wiki.toml — per-wiki config (committed to git, shared)

Config file location

1. --config <path> CLI flag (highest)
2. LLM_WIKI_CONFIG env var
3. ~/.llm-wiki/config.toml (default)

The --config flag is global — it works on every subcommand including serve. Use it in .mcp.json to run isolated environments:

json
{
"llm-wiki": {
"command": "llm-wiki",
"args": ["--config", "/path/to/config.toml", "serve"]
}
}

Or via env:

json
{
"llm-wiki": {
"command": "llm-wiki",
"args": ["serve"],
"env": { "LLM_WIKI_CONFIG": "/path/to/config.toml" }
}
}

Resolution order (settings within the config)

1. CLI flag / MCP parameter
2. Per-wiki config (wiki.toml)
3. Global config (config.toml)
4. Built-in default

Per-wiki wins over global. CLI flags win over everything.

wiki.toml

Lives at the wiki repository root. Contains wiki identity and optional per-wiki overrides.

toml
name = "research"
description = "ML research knowledge base"
[ingest]
auto_commit = true
[validation]
type_strictness = "loose"
FieldRequiredDescription
nameyesWiki name — used in wiki:// URIs
descriptionnoOne-line description

Type overrides ([types.*]) are also declared here but managed through the schema skill.

List current config

wiki_config(action: "list")

Returns all resolved config values for the default wiki.

json
{
"defaults.search_top_k": 10,
"defaults.page_mode": "flat",
"ingest.auto_commit": true,
"validation.type_strictness": "loose"
}

Read a config value

wiki_config(action: "get", key: "defaults.search_top_k")

Set a config value

Per-wiki (writes to wiki.toml):

wiki_config(action: "set", key: "ingest.auto_commit", value: "false")

Global (writes to config.toml):

wiki_config(action: "set", key: "defaults.search_top_k", value: "15", global: true)

Global-only keys reject per-wiki writes with an error.

Overridable settings

These keys can appear in both config.toml and wiki.toml.

KeyDefaultDescription
defaults.search_top_k10Default result count for search
defaults.search_excerpttrueInclude excerpts in search results
defaults.search_sectionsfalseInclude section pages in search
defaults.page_modeflatDefault page creation mode: flat or bundle
defaults.list_page_size20Default page size for list
defaults.output_formattextDefault output format: text or json
read.no_frontmatterfalseStrip frontmatter from read output
ingest.auto_committrueCommit to git after ingest
validation.type_strictnessloosestrict: unknown type is error; loose: warning
graph.formatmermaidGraph output format: mermaid, dot, json, or llms
graph.depth3Default hop limit for graph traversal
graph.type[]Page types to include; empty = all
graph.output""Default output path; empty = stdout
graph.snapshottruePersist graph to disk for warm-start across restarts
graph.snapshot_formatbincode+lz4Snapshot serialization format: bincode+lz4 or bincode+zstd
graph.structural_algorithmstrueEnable diameter/radius/center in wiki_stats; false skips O(n²)
graph.max_nodes_for_diameter2000Skip structural algorithms when local node count exceeds this
graph.min_nodes_for_communities30Minimum pages before Louvain community detection runs
search.status.<key>—Override search weight for a status value; key is a status slug (e.g. active, draft), value is a float
index.memory_budget_mb50Tantivy writer memory budget in MB
index.tokenizeren_stemTantivy tokenizer for text fields

Search status weights

Control how much each status value contributes to search ranking. Higher weight = pages with that status rank higher in results.

wiki_config(action: "set", key: "search.status.active", value: "1.2")
wiki_config(action: "set", key: "search.status.draft", value: "0.3")
wiki_config(action: "set", key: "search.status.archived", value: "0.1")

Custom status values defined in your schemas work the same way:

wiki_config(action: "set", key: "search.status.accepted", value: "1.5")

Per-wiki overrides take precedence over global values — use them when a wiki has a domain-specific status taxonomy (e.g. ADR accepted/superseded).

Global-only settings

These keys can only appear in config.toml. Setting them in wiki.toml is rejected.

KeyDefaultDescription
index.auto_rebuildfalseRebuild stale index before search/list
index.auto_recoverytrueRebuild corrupt index on open failure
serve.httpfalseEnable HTTP transport
serve.http_port8080HTTP port
serve.http_allowed_hostslocalhost,127.0.0.1,::1Allowed Host headers
serve.acpfalseEnable ACP transport
serve.max_restarts10Max transport restarts; 0 = no restart
serve.restart_backoff1Initial backoff seconds; doubles, cap 30s
serve.heartbeat_secs60Heartbeat interval; 0 = disabled
serve.acp_max_sessions20Max concurrent ACP sessions; NewSession returns error when exceeded
logging.log_path~/.llm-wiki/logsLog file directory; empty = stderr only
logging.log_rotationdailydaily, hourly, never
logging.log_max_files7Max rotated files; 0 = unlimited
logging.log_formattexttext or json
watch.debounce_ms500Filesystem watcher debounce interval in ms

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/geronimo-iia/llm-wiki-skills/config/badge)](https://verified-skill.com/skills/geronimo-iia/llm-wiki-skills/config)

Something wrong with this skill or its page? Tell us.