RepoBridge

Local-first GitHub repo inspector for grounded LLM context.


Diagnostics completed

karpathy/nanoGPT · run 36eec7b0-9f2d-4382-a1aa-902b04420262 · 2026-08-26 10:40 UTC

This run is partial. Every stage completed, but at least one was bounded — a limit, a budget, or a truncated listing. Each is quantified below.

8 info 1 partial 0 degraded 0 error

Stages

In pipeline order. A stage that never ran is absent rather than shown as healthy — for a run that failed early, that absence is the point.

validation 1 metadata 1 tree 1 classification 2 content 1 extraction 1 summarization 1 persistence 1

Level: info partial degraded error

Timeline

9 events, in the order the pipeline produced them.

  1. info validation 10:40:22

    Repository URL accepted as karpathy/nanoGPT.

    owner
    karpathy
    name
    nanoGPT
  2. info metadata 10:40:22

    Repository metadata retrieved for karpathy/nanoGPT.

    default_branch
    master
    language
    Python
    archived
    False
    fork
    False
  3. info tree 10:40:22

    Tree retrieved for ref 'master': 32 entries.

    resolved_ref
    master
    entries
    32
  4. info classification 10:40:22

    Classified 32 entries: 24 selected, 2 skipped, 6 directories.

    selected
    24
    skipped
    2
    failed
    0
    directories
    6
    accounts_for_every_entry
    True
  5. partial classification 10:40:22

    2 entries were skipped without reading. Every one is listed with its reason in the file view.

    skipped_non_text_asset
    2
  6. info content 10:40:23

    Read 24 of 24 selected files.

    files_read
    24
    selected
    24
    read_cap
    100
    concurrency
    8
    accounts_for_every_selected_file
    True
  7. info extraction 10:40:23

    Extracted 65 chunks from 24 files.

    chunks
    65
    files
    24
  8. info persistence 10:40:23

    Run stored with 32 classified entries and 65 chunks.

    run_id
    36eec7b0-9f2d-4382-a1aa-902b04420262
    entries
    32
    chunks
    65
  9. info summarization 10:40:31

    Summary generated by gemini-3.5-flash-lite.

    model
    gemini-3.5-flash-lite
    prompt_tokens_estimated
    22421
    chunks_included
    65
    chunks_total
    65

Diagnostics is an account of the run, not a second copy of it. Per-file detail for every entry — including everything skipped without reading — is in the file listing, which is why a repository with 22,924 skipped files still has a readable timeline here.