AGENTS.md

1# @second-brain/collectors — Agent Instructions
2 
3Data collectors and streaming providers that feed the knowledge graph.
4 
5## Entry Point
6 
7`src/pipeline/runner.ts` → `PipelineRunner` with `register()` and `run()`.
8 
9## Collectors (6)
10 
11| Collector | Directory | What it collects |
12|-----------|-----------|-----------------|
13| `git` | `src/git/` | Commits, branches, file history |
14| `ast` | `src/ast/` | Code symbols, dependencies, structure |
15| `github` | `src/github/` | Issues, PRs, reviews, comments |
16| `gitlab` | `src/gitlab/` | MRs, issues, pipelines |
17| `conversation` | `src/conversation/` | Chat/conversation transcripts |
18| `docs` | `src/docs/` | Documentation files (markdown, etc.) |
19 
20## Providers (5)
21 
22| Provider | Directory | Transport |
23|----------|-----------|-----------|
24| `custom` | `src/providers/` | Programmatic API |
25| `git` | `src/providers/` | Git hooks / polling |
26| `github` | `src/providers/` | GitHub webhooks |
27| `gitlab` | `src/providers/` | GitLab webhooks |
28| `webhook-relay` | `src/providers/` | Generic webhook relay |
29 
30## Adding a New Collector
31 
321. Implement the `Collector` interface in a new `src/<name>/` directory
332. Register in `src/pipeline/runner.ts`
34 
35## Other Directories
36 
37- `src/extraction/` — LLM-based entity extraction helpers
38- `src/git-context/` — Git context resolution
39- `src/realtime/` — Real-time event streaming
40- `src/watch/` — File system watchers
41