Skills Catalog
Skills are auto-triggered context providers. They activate based on natural-language signals in the conversation — no explicit invocation needed. Each skill provides guidance and suggests the most relevant command for the current moment.
How Skills Work
Skills listen for keywords and patterns in the conversation. When a match is detected, the skill surfaces a suggestion in the conversation context. The user can act on the suggestion or ignore it — skills never execute commands automatically.
Skill Reference
| Skill | Triggers On | Suggests |
|---|---|---|
| lesson-capture | "figured it out", "solved it", "pattern here is", bug breakthrough, debugging victory | /kmgraph:capture-lesson with pre-filled context |
| kg-recall | "have we done this before", "did we solve this", past decisions, history questions | /kmgraph:recall with extracted search terms |
| session-wrap | Session ending, context limit approaching, major milestone completed | /kmgraph:session-summary before compaction |
| adr-guide | "I'm thinking of using...", architecture decisions, tech choice discussions | /kmgraph:create-adr with decision guidance |
| doc-update-router | "update [doc name]", "update the session summary", "update the changelog" | Routes to the correct update command; bypasses direct file edits |
| stuck-work-escalation | A bug, test, or task has resisted 3+ distinct fix attempts or 30+ minutes of effort | Activates Opus diagnosis gate, enforces hypothesis logging per attempt, drives to structured exit-path decision at 5 attempts |
| docs-impact-scan | "push to origin", "push and merge", "push and merge with admin", "open PR", "create PR", "finishing up", "ready to push" | Scans for all user-facing docs affected by current branch changes; validates list with user before dispatching /kmgraph:update-doc --user-facing for each confirmed file |
| sidebar-update | A docs file is moved or renamed, "move [doc]", "rename [doc]", git mv on a docs/ path | Finds the stale id: entry in sidebars.js, updates it to the new path, scans for broken internal links |
| capture-router | "capture that", "remember that", "save that", and similar natural-language capture phrases | Auto-detects content type and destination from content signals; presents single confirmation before writing |
| rules-capture | "always X", "never X", "from now on X", "don't do X" (standing rule), "I prefer X", "make that a rule" — implicit behavioral corrections without capture vocabulary | Detects scope (project vs personal, rule vs me), shows one-line suggestion with 4-target shortcut menu, dispatches to rules-capture-agent on confirmation |
| gov-execute-plan | "execute plan", implementation start, docs/plans/*.md mentioned | Zero-deviation 8-step execution protocol |
| knowledge-graph-usage | Questions about KMGraph itself, "how do I...", "what command..." | Orients to the four-layer architecture and surfaces relevant commands |
Trigger Keywords
Skills use pattern matching on natural language, not exact commands. The lesson-capture skill fires on phrases that indicate a breakthrough:
- "figured it out"
- "solved it"
- "the pattern here is"
- "the fix was"
- "turns out the issue was"
- "what I learned"
The capture-router skill fires on capture intent phrases:
- "capture that"
- "remember that"
- "save that"
- "let's document this"
Skills are designed to catch capture opportunities that users would otherwise miss because they are in flow.
Skills vs Commands
Skills provide suggestions. Commands execute actions. A skill surfaces /kmgraph:capture-lesson at the right moment; the command does the actual capture. Skills can be thought of as a smart prompt layer that knows when to interrupt.
Related
- Agents Catalog — Heavy-lift handlers triggered by skills
- Command Guide — Full command reference
- Concepts — Automation Layer — How skills, agents, and hooks fit together