| name | find-connections-a |
| type | action |
| description | This skill should be used when the user asks to "find connections", "what relates to this", "how does this connect", "map connections", "show related work", or needs to discover relationships across the vault. |
| argument-hint | <filepath | topic | area-name> |
Action skill — Connection discovery: semantic search, backlink analysis, tag-based discovery, PARA structure analysis.
Connection Discovery
Combines semantic search, backlinks, tags, and PARA structure to map relationships across the vault.
Input Parsing
The argument can be:
- File path → Read the note, use its content as the search basis
- Topic text (e.g., "prediction market pricing") → Use directly as search query
- Area name (e.g., "ai-dev-ecosystem") → Read the area's L0 MOC, use its thesis and key concepts
Process
1. Semantic Search
Query the vault for similar notes.
For a note or topic:
qmd query "<key sentences or topic>" --files
For an area, run multiple queries based on the MOC's sections:
qmd vsearch "<section 1 topic>" --files
qmd vsearch "<section 2 topic>" --files
Collect top 10-15 results across queries.
2. Backlink Analysis
- Find all notes that link to/from the target
- Trace 2-hop connections: notes that share common links but don't directly link to the target
- Use Grep: search for
[[note-name]] patterns across the vault
3. Tag-Based Discovery
Find notes with knowledge-role tags that suggest conceptual relationships:
#seed + #seed in similar areas → potential threads waiting to be connected#question notes → might pose questions the target answers#tool notes → frameworks that apply to the target's domain#insight notes → crystallized understanding that builds on the target- Shared tags + semantic similarity = strong connection signal
4. PARA Structure Analysis
Check structural relationships:
- Which projects draw from the same area?
- Which areas have overlapping scope with the target?
- Which resources were captured for the same purpose?
- Read project briefs to find area MOC references
5. Synthesize Connection Map
Present findings organized by relationship type:
Direct connections — linked or semantically close:
- [[2025-01-21_agentic-primitives]] (ai-dev-ecosystem)
Semantic similarity: high | Shared concepts: agent architecture, tooling
Indirect connections — shared links, related tags, related area:
- [[2025-02-01_transformer-scaling]] (modeling)
Connection via: both linked from [[00_ai-dev-ecosystem]], both tagged #thread
Project opportunities — area knowledge that could feed a project:
- prediction-markets area notes → prediction-market-algo-dev project
Notes about market mechanics could inform algorithm design
Gaps — topics mentioned but with no notes:
- "memory continuity" mentioned in 3 notes but no dedicated note exists
- "cross-market correlation" referenced in project brief but no area coverage
Output Format
## Connection Map: {input}
### Direct (N notes)
{strongest connections}
### Indirect (N notes)
{2-hop links, shared tags, related areas}
### Project Opportunities
{area notes that could serve projects}
### Gaps
{topics referenced but not covered}
When to Use vs. Other Skills
- backlinks-a — focused on one note's links. Use when you know the note.
- find-connections-a — broad exploration across the vault. Use when mapping a topic or area.
- sem-search-a — raw search results. Use when you just need to find notes.
