typemd

This skill should be used when the user asks to "resolve an issue", "work on issue #N", "fix #N", "implement #N", "close #N", "tackle #N", "pick up #N", "start working on #N", "what should I work on next", or references a specific GitHub issue number they want to work on. Can also accept a version number (e.g., "0.5.0") to select from that release's sub-issues, or auto-select the best issue when no argument is specified.

Use when a typemd code change needs end-to-end TUI verification — keybinding behaviour, layout rendering, config overrides, error recovery — and "open TTY failed" or "non-interactive sandbox" is about to be treated as a blocker. Trigger on symptoms like "can't test TUI without a terminal", "need to verify what the user sees", "press ctrl+X and check X appears", "help popup should show Y", "sidebar should look like Z after config change".

Import existing markdown collections into a structured typemd vault. Use when the user wants to onboard, import, or migrate content from external directories into their vault.

Import existing markdown collections into a typemd vault.

This skill should be used when the user asks to "create an issue", "open an issue", "file a bug", "request a feature", "add a task", or discusses a problem or idea that should be tracked as a GitHub issue.

Use when resolving a GitHub issue labeled `discussion` — facilitates decision-making on open questions, documents conclusions as an issue comment, creates follow-up issues if needed, and removes the discussion label. Triggered by resolve-issue when it detects a discussion-labeled issue, or directly when user asks to "resolve discussion

Cross-reference BDD tests, OpenSpec specs, and documentation to find contradictions and terminology inconsistencies, then optionally triage coverage gaps by verifying Go source code. Use when the user mentions "consistency check", "cross-reference", "do BDD and spec align", "are docs and tests contradicting", "is terminology consistent", "check specs", "spec and feature match", "triage coverage gaps", "verify implementation matches spec". Also useful before releases, after refactoring, or when suspecting a concept is described differently across sources.

Use when reviewing BDD feature files or OpenSpec specs for technical-only scenarios without user context, when cleaning up test layers, when deciding whether a behavior belongs in BDD or unit tests, or when the user mentions "check features", "BDD cleanup", "BDD audit", "feature audit", "test layer review".

Use when preparing a release, writing release notes, creating a GitHub draft release, or updating CHANGELOG.md. Triggers on "write release notes", "prepare release", "create release", "update changelog".

Explore existing markdown files and suggest typemd type schemas and properties. Use when the user wants to analyze their files before importing, plan their vault structure, or get type/property recommendations. Read-only — never creates, modifies, or deletes any files.

Convert existing markdown files into typemd objects. Use when the user wants to import, migrate, or convert markdown files into their typemd vault.

typemd vault reference. Use when designing type schemas, creating objects, managing relations, or working with vault structure — provides CLI commands, type schema format, object format, wiki-links, views, templates, and TUI keybindings.

Analyze existing markdown files and suggest typemd type schemas.

Convert existing markdown files into typemd objects.

Guide for using tmd instructions to output embedded skills enriched with vault context. Use when integrating typemd with AI tools, passing vault context to LLMs, or setting up skill-based workflows.

Use when designing knowledge organization features — type schemas, relations, labeling, hierarchy, taxonomy, tagging, metadata, or cross-vault mapping. Use when the user mentions "SKOS", "Dublin Core", "Schema.org", "taxonomy", "concept hierarchy", "broader/narrower", "preferred label", "alias", "metadata standard", "knowledge organization", or when evaluating whether a data model aligns with established knowledge management standards.

Use when calculating panel widths, heights, or split layouts in the TUI using lipgloss v2. Use when panels don't fill the screen, content overflows borders, or split panels have gaps. Triggered by "Width() semantics", "border width calculation", "panel doesn't fill", "content overflows panel".

Use when the user asks to "plan a roadmap", "review release scope", "rebalance releases", "check if release is too large", or wants to adjust issue distribution across versions to fit a one-week release cadence.

Retrospective analysis and correction of skills based on execution experience. Analyzes the current conversation to identify what went wrong during skill execution and applies targeted fixes to the SKILL.md and metadata.

Use when the user asks to "break down an epic", "split an epic into sub-issues", "create sub-issues for an epic", or wants to decompose a large epic issue into smaller, actionable sub-issues on GitHub.

Use when the user wants to write a blog post — technical articles, devlogs, opinion pieces, or any non-release blog content. Triggers on "write a blog post", "write an article", "create blog post". Do NOT use for release notes — those are handled by the `create-release` skill.

Use when code has changed and documentation may be outdated, before releases, after adding new features or commands, or when user asks to check if docs match implementation. Symptoms include new commands without reference pages, changed UI without updated screenshots or descriptions, missing CLI flags in docs.

Use when blog posts have been added or updated in zh-tw, when user asks to sync or translate blog content to English, or before releases to ensure both language versions are in sync.