AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Wiki Save

skill-santiagobobrik-llm-wiki-kit-wiki-save · by SantiagoBobrik

Save a conversation answer or insight as a Wiki article. Use when the user says 'save this to wiki', 'wiki-save', or wants to file a good answer into the knowledge base.

No reviews yet
0 installs
7 views
0.0% view→install

Install

$ agentstack add skill-santiagobobrik-llm-wiki-kit-wiki-save

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution Used

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-santiagobobrik-llm-wiki-kit-wiki-save)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
4mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Wiki Save? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

wiki-save

Saves conversation content as a structured Wiki article. This is the "query → file back" operation — when a conversation produces a good answer, analysis, or synthesis, this skill files it into Wiki/ so it compounds in the knowledge base instead of disappearing into chat history.

Arguments — parse $ARGUMENTS:

  • --vault or --vault= — explicit vault. Pass . to mean "the vault at cwd" (resolved against the pre-loaded list).
  • Everything else in $ARGUMENTS, after removing the flag and its value, is the topic — the article title/topic. If empty, infer from the conversation context.

Vault resolution (only when --vault is absent):

  • Read the pre-loaded vault list:
  • Exactly one vault → use it. Do not ask.
  • Two or more → ask via AskUserQuestion.
  • None → bail; Obsidian isn't running.

--vault . handling: resolve the current working directory against the pre-loaded vault list. If cwd matches a registered vault path, use that vault's name. Otherwise bail — cwd isn't a vault.

Pre-loaded Context

Registered Obsidian vaults (namepath per line):

!obsidian vaults verbose 2>/dev/null

Workflow

0. Select the target vault

Parse the pre-loaded vault list (namepath per line). Apply the Vault rule from the top of this file to pick one. Hold the chosen vault's ` and absolute ` — every subsequent step uses them.

If /Wiki/ or /Wiki/index.md is missing, print ℹ️ Vault not wiki-initialized. Run /wiki-init to bootstrap. and stop.

1. Identify the content to save

Look at the recent conversation. The content to save is typically:

  • An answer to a complex question the user asked
  • An analysis or comparison the user found valuable
  • A synthesis that connects multiple ideas
  • A decision or conclusion reached during discussion

If it's unclear what to save, ask the user.

2. Determine placement

Read the chosen vault's index (cat /Wiki/index.md) and check for existing articles on the topic:

  • If a related article exists → ask the user: update the existing article or create a new one?
  • If no related article exists → create a new one

Pick the right subfolder (Wiki//) using the same categories as wiki-compile (AI, Engineering, Business, Books, etc.).

3. Write the article

Structure the conversation content into a clean article. Not a transcript dump — synthesize it into something readable.

Read the article template from ${CLAUDE_SKILL_DIR}/assets/article-template.md and fill in the placeholders. That file is the source of truth — edit it to change the default article shape for saved articles.

Key rules:

  • Always use [[wikilinks]] — every reference to a Wiki article or Raw clipping must be a wikilink. No plain paths.
  • origin: conversation in frontmatter — distinguishes articles born from Q&A vs compiled from Raw sources.
  • Synthesize, don't transcript — the article should read as a standalone piece, not as "I asked X and you said Y."
  • Link related Wiki articles inline where relevant.
  • Granular type — pick from concept, guide, pattern, analysis, comparison, reference rather than generic notes. Add a new one only if none fit.
  • Preserve specifics — concrete examples, numbers, code patterns, and named tools/frameworks. Before closing the article, check against the conversation: did the named people / tools / companies survive? The numbers and dates? The concrete examples (commands, snippets, workflows)? The tensions or disagreements? If the article could have been written from the title alone, rewrite it.
Enriching vs splitting

If a related Wiki article already exists, the default pull is to append. Resist it when the new material is actually a different concept wearing the same category.

Split into a new article (and wikilink from the old one) when:

  • You're about to open a third top-level ## section about a sub-topic that wasn't in the original scope.
  • The sub-topic has its own vocabulary, its own named tools, its own sources — it could stand alone.
  • You can't summarize the resulting article in one sentence without using "and" to join two unrelated concepts.
  • A reader looking only for the sub-topic would have to scroll past unrelated material to find it.

Enrich the existing article (don't split) when the new material adds depth to what's already there: more examples, sharper definitions, a counter-argument, a concrete number.

Tone

Wiki articles are read, not sold. Reference entries, not blog posts. Flat, factual, present tense. The reader came for the facts — don't perform, don't hype, don't narrate.

Avoid:

  • Hype adjectives: "powerful", "revolutionary", "game-changing", "seminal", "elegant", "robust".
  • Editorial voice: "interestingly", "notably", "it's worth noting", "crucially", "importantly".
  • Progressive narrative: "would later become", "eventually led to", "goes on to".
  • Empty qualifiers: "genuine", "profound", "deep", "meaningful".
  • Decorative em dashes (—). Use a period, a comma, or parentheses. Reserve em dashes for real asides.

Prefer short affirmative sentences. When a phrase from the conversation carries weight, quote it directly — the quote holds the emotion, the surrounding prose stays neutral.

4. Create via obsidian CLI

obsidian vault= create path="Wiki//.md" content="" silent

Then force date types:

obsidian vault= property:set name="created" value="YYYY-MM-DD" type=date path="Wiki//.md"
obsidian vault= property:set name="updated" value="YYYY-MM-DD" type=date path="Wiki//.md"

5. Update index

Add the new article to the chosen vault's Wiki/index.md under the appropriate category heading:

- [[Wiki//|]] — 

6. Confirm

Tell the user what was saved and where:

Saved to Wiki/AI/llm-wiki-pattern.md
Updated Wiki/index.md.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.