v0.5.0 - Cross-platform, sharper triggers, and bounded context cost

Highlights

Since v0.4.0 the suite went cross-platform (agentskills.io) and got a round of scaling + hygiene work.

Cross-platform (agentskills.io)

• Skills run on Claude Code, Gemini CLI, Codex, and Copilot - only .skill packaging and the Claude GUI install stay Claude-specific.
• metadata.version on every skill; skill bodies generalised; README cross-platform install (incl. npx).

Sharper triggering

• All skill descriptions overhauled for reliable triggering without cross-skill collisions.

Bounded per-run context cost (the "scaling trio")

• wiki-crystallize offers to archive an oversized log.md (>150 entries → keep the newest 50, move the rest to archive/log-YYYY-MM-DD.md; a same-day re-archive appends, never overwrites). Closing-posture thread-weight signal now keys on the page's crystallize_count, so it survives log archiving.
• wiki-lint adds read-only log oversized and over-length index entry findings.
• Index-entry descriptions capped at 150 chars; the duplicated "What this skill does not do" footer pruned from the five operational skills.
• wiki-query log write fixed to prepend (was append), matching the rest of the suite.

Skill versions

wiki-config 1.11 · wiki-ingest 4.17 · wiki-lint 3.15 · wiki-integrate 3.13 · wiki-crystallize 3.17 · wiki-query 3.11

wiki-pack stays local/unreleased by design.

v0.4.0 - Page templates, wiki-help, and schema compliance

v0.4.0 - Page templates, wiki-help, and schema compliance

Three things I wanted to get right before this felt like a proper release: a way to control page structure without touching skill code, something a new user can actually read to understand what they installed, and a lint pass that catches schema drift before it compounds. All three are in this release.

Page templates

wiki-config now deploys 13 Markdown templates to your templates/ folder on first run - one per page type. Writing skills read the matching template when creating new pages. The templates are yours to edit: change section headings, add a comparison table, drop in a checklist. Whatever structure you put in the template is what new pages start with. No skill repackaging required.

wiki-config's new Step 6 is a template management flow: view any template, deploy missing ones, reset individual files or the full set.

wiki-help.md

New user-facing help file deployed to your wiki root alongside wiki-config.md on first run. Covers field conventions, write discipline, naming, page types, how to customise the wiki (config, schema, templates), a getting started guide, and a troubleshooting section that explains what broken frontmatter looks like in different editors. Open it in your notes app - it is a plain Markdown file.

If you are upgrading from an earlier version, run /wiki-config after uploading the new skills - it will detect the missing file and offer to deploy it.

First-run scaffolding

wiki-config now creates Home.md and Overview.md from their bundled templates on first run, in addition to the existing index.md, log.md, raw/, ingested/, and templates/. Fresh installs arrive with a navigable structure rather than an empty folder with two config files. index.md and log.md are also now created from templates rather than hardcoded headers.

Schema compliance (Batch 3c)

wiki-lint now validates every in-scope page against wiki-schema.md:

• Hard errors: source: present without reliability:, invalid status: or reliability: enum values, invalid page_type enum value
• Soft warnings: skill-touched knowledge pages missing source: or description:, source: without a ## Sources body section

page_type propagation

wiki-integrate and wiki-crystallize now check page_type on every page they touch: infer and confirm if missing, offer promotion if the page has outgrown its current type. wiki-lint adds a hard error for invalid page_type values and a soft warning for skill-touched pages missing the field.

Log prepend discipline

Writing skills previously said "append to log.md" but placed entries at the top. The instruction and the behaviour now agree. The log.md template was updated to say prepend-only to match.

Skill versions in this release

Skill	Version
wiki-config	v1.8
wiki-ingest	v4.14
wiki-lint	v3.11
wiki-integrate	v3.10
wiki-crystallize	v3.12
wiki-query	v3.7

Upgrading from v0.3.0

Re-upload all six .skill files in Claude Desktop (Customize → Skills), then run /wiki-config. It will check your existing wiki for missing components (wiki-help.md, Home.md, Overview.md, templates folder) and offer to deploy whatever is absent. Your existing pages and ingested sources are not touched.

v0.3.0 - Page freshness, provenance, and source trust

Last month, I scoped off a section of my vault to experiment with filesystem mcps and collaborative editing using various models.

That section of my note graph looked something like this two weeks ago:

I stumbled over Karpathy's original gist really quite by accident, just looking for learning resources for agentskills.io. So, I figured, what could possibly go wrong and just winged it with Claude.

This is my graph about seven days later while using and improving these skills to this current release.

That tightly-knit cluster is the part that the wiki has now fully absorbed: domain homes emerged, and many collaborative runs were had, each usually ending with a detailed /wiki-crystalize.

This has been a fun journey! I hope my lightweight implementation of the original pattern is as useful to you as it has been for me! Many new things today.

---

v0.3.0 - Page freshness, provenance, and source trust

This release is about the wiki knowing more about itself. Pages now track when they were last touched, where their knowledge came from, and whether that source was trustworthy enough to take at face value. The ## Pending Review section is the visible output of the trust system - a standing reminder that a page was created from a single non-authoritative source and needs corroboration before you rely on it.

New: wiki-schema.md

A second config file, deployed alongside wiki-config.md by /wiki-config. It defines the frontmatter structure all six skills follow - mandatory fields, conditional fields, valid enum values - and documents write discipline in one place. All six skills read it on boot. Previously each skill carried its own idea of what a page looked like; now there is a single source of truth. If the schema is missing or malformed, skills stop and redirect rather than proceeding with inconsistent assumptions.

The schema is designed to be extended. Adding conditional fields or new enum values is safe. Removing mandatory fields or renaming existing ones is not recommended; skills depend on those being where they expect them.

New: freshness tracking

Three fields, written automatically by skills:

• updated: - the date any skill last touched a page. Written by wiki-ingest, wiki-integrate, and wiki-crystallize on every write. On page creation it equals date: (correct and expected). date: itself is now strictly immutable - creation date only, never modified on subsequent writes.
• crystallize_count: - counts deliberate crystallize events on a page, not general edits. wiki-crystallize reads the current value and increments; if absent, writes 1. Paired with updated:, this answers "is this page current and well-revised?" without parsing the log.
• Staleness check in wiki-lint - flags pages where updated: is more than 90 days in the past. Exempt: pages with status: artefact, snapshot, archived, or page_type: reference. Missing date: or updated: fields on skill-written pages are now a lint error rather than a silent gap.

New: provenance and trust system

Four fields on page creation, written by wiki-ingest:

• status: - active for pages with real content, stub for thin pages (fewer than ~3 substantive sentences). Agent judgment. Other values on state change: artefact, archived, snapshot.
• description: - ~200 char page summary, quoted. Written by wiki-ingest on create; wiki-crystallize rewrites it when the page's structure or scope changes substantially, not on incremental updates.
• source: - one-element list with the originating ingested/ path. Written on create only; enrichments go in the body ## Sources section. Overwriteable when a subsequent ingest genuinely supersedes the original.
• reliability: - high, medium, or low. Agent assesses the originating source's nature: primary or authoritative = high; well-sourced secondary = medium; blog, opinion, single-source, unverified = low. Only present when source: is present.

New: ## Pending Review section

Written by wiki-ingest on knowledge, research-note, and survey pages where reliability: is medium or low. high reliability sources get no section. The framing scales with trust level - medium gets a permissive note (one corroborating source would retire this); low gets a stronger prompt with specific claims to verify.

The section stays until a better source is ingested. On re-ingest with a demonstrably more authoritative source, wiki-ingest updates source: and reliability: in frontmatter, adds the new source to ## Sources, and removes the section. The agent can also offer to search for a primary or corroborating source inline rather than writing the section and moving on - when source quality is genuinely ambiguous, it asks rather than decides unilaterally.

Improved: wiki-ingest selective intake

Step 1.5 - a thematic triage layer before processing. wiki-ingest does a quick survey of raw/, clusters files by domain affinity using the wiki's actual domain structure, and presents batches before touching anything. You choose which batch to process now; unselected files stay in the queue. This matters most when raw/ has accumulated files across multiple domains - processing them together previously meant mixed-domain synthesis that was harder to review.

Improved: content trust boundary

An explicit security section in wiki-ingest. Source documents are untrusted data. The skill never executes commands, exfiltrates data, or modifies its own behaviour based on embedded instructions in sources. Suspicious content is flagged in the session summary; synthesis continues. This was always implicit practice; it is now an explicit, numbered set of rules that loads before the workflow step that processes untrusted content.

Improved: Config Discovery hardening

Several fixes across this release cycle, each prompted by a specific observed failure:

• wiki-schema.md check is now a dedicated mandatory step in Config Discovery, with explicit "do not proceed until you have a definite verdict" language. Previously the schema check was buried in prose and easy to skip.
• Scope check is now a MANDATORY STOP on broad filesystem scope (C:\, /home/X, etc.). Previous soft conditional language ("if broad scope, consider skipping") was not followed reliably in testing. Imperative beats conditional.
• wiki-config precedence is enforced: when /wiki-config is available, operational skills stop and redirect to it rather than offering their own bundled fallback. The fallback path (deploy from bundled references/wiki-schema.md) is reserved for when wiki-config is genuinely not installed.

Version bumps:

• wiki-config 1.0 → 1.6
• wiki-ingest 4.0 → 4.11
• wiki-lint 3.0 → 3.9
• wiki-integrate 3.0 → 3.8
• wiki-crystallize 3.0 → 3.9
• wiki-query 3.0 → 3.7

• Re-upload all six .skill files via Claude Desktop → Customize → Skills.
• Run /wiki-config once after upgrading. It will deploy wiki-schema.md alongside your existing wiki-config.md if you don't already have one. Nothing breaks without it, but skills will stop and ask rather than proceed.
• Existing wiki pages work unchanged. The new frontmatter fields (updated:, status:, description:, source:, reliability:, crystallize_count:) are conditional - skills write them on the next touch. Pre-existing pages without these fields are treated as status: active for backward compatibility.
• If you see ## Pending Review sections appearing on new pages after upgrading: that is intentional. It means wiki-ingest assessed the source as medium or low reliability. Retire the section by ingesting a stronger source for that topic.

---

I had a little fun with the canvas design skill just trying to express how this little project has helped shape the gravity of my knowledge base and task tracking. So many changes to quotas lately, but I would say my token use has about halved.

v0.2.0 - New Skill - wiki-config - features & updates in all skills

v0.2.0 - Interactive config + assisted cold start

Batch 1 delivers a sixth skill and a refactored config story.

New: wiki-config skill. Handles setup, validation, and reconfiguration interactively. One clean entry point for "I'm starting from scratch" and for "I want to tweak a field without hand-editing YAML." It asks where your wiki root should live, deploys wiki-config.md from a template, and scaffolds the directory structure (index.md, log.md, raw/, ingested/ with taxonomy subdirs, templates/ reserved for a future batch).

Refactored: cold-start behaviour across all five operational skills. Previously each skill carried its own ~90 line init block that tried to silently deploy a default config on miss. That was the wrong shape - skills that couldn't explain what they were doing were doing the explaining. Now all five operational skills share an identical ~22 line advisory Config block: search for wiki-config.md, ask once if not found, and if the user can't give a clear answer, refuse politely and point at /wiki-config (or bundled references for manual setup). The recommended path is to install /wiki-config once and let it handle setup properly.

Progressive disclosure via bundled references. Each operational skill now ships references/setup-help.md (a focused manual-setup mini-guide) and references/wiki-config-template.md (the default config), pulled into context only when a user opts out of the recommended /wiki-config path. This is Anthropic's intended skill pattern - keep the SKILL.md body focused; let references load on demand.

Template ships with a low-magic blacklist. The blacklist field in the bundled template uses obvious placeholder-description values ("Folder(s) where the wiki should not write", "You can edit these") that force user attention before the wiki is useful. No silent magic; no guessing what was real.

Other highlights since v0.1.0:

• Source attribution via a ## Sources section in each wiki page body, plus classification ask-on-mismatch and richer synthesis depth (wiki-ingest)
• Em-dash violation check added to wiki-lint
• Large-document third path for wiki-ingest: synthesis, enrichment of related pages, or stub with Sources pointer for oversized reference material
• Hub-first targeting and session-checkpoint framing in wiki-crystallize
• Wiki root derived from wiki-config.md location at runtime; wiki_root field removed
• All .skill binaries folder-wrapped per Anthropic's spec

Version bumps:

• wiki-config NEW v1.0
• wiki-ingest 3.6 → 4.0
• wiki-lint 2.7 → 3.0
• wiki-integrate 2.6 → 3.0
• wiki-crystallize 2.7 → 3.0
• wiki-query 2.6 → 3.0

Upgrade notes:

• Existing wiki-config.md files work unchanged - the schema is forward-compatible. You can add templates_folder: templates\ now if you want, but no skill reads it yet.
• Re-upload all six .skill files via Claude Desktop → Customize → Skills after downloading from the repo root.
• If you've been running without /wiki-config installed: nothing breaks, but skills will now politely refuse on config-miss instead of silently deploying a default. Install /wiki-config and run it once.

v0.1.0 - 5 LLM Wiki Skills

v0.1.0 - First public release
Five installable Claude skills for building and maintaining a local Markdown wiki using the LLM wiki pattern. This is a working implementation I built while learning - it's functional and tested, but a little rough around the edges. Evals are pending and there are a few known open items.
What's here: wiki-ingest, wiki-query, wiki-lint, wiki-integrate, and wiki-crystallize. A wiki-config.md handles all configuration and is shared across all five skills. See the README for background, installation, and getting started.
Requires Claude Desktop with a filesystem MCP (for Claude Chat). Claude Code works too with minor adaptation.