Tool · Hermes · Nous Research
read_file
How this differs from most implementations
- Staleness check before writeinvariantholdsvs fails in 6 of 8
- Byte / char cap100k chars (not lines)vs 50 KB in 3 of 5
- Whichever hits firstcharsvs lines or bytes in 4 of 5
- Giant single linechar budget absorbs itvs per-line cap in 4 of 6
Derived by comparing every recorded value against the other implementations of the same capability. Differing is not a defect unless the row is marked an invariant.
Observed · Read
compare all →| Output shape | |
| Dedicated read tool | read_filesrc |
| Line numbering | N|src |
| Output wrapper | plainsrc |
| Limits | |
| Default line cap | 2000src |
| Byte / char cap | 100k chars (not lines)src |
| Per-line cap | — |
| Whichever hits first | charssrc |
| Pagination | |
| Offset base | 1-indexedsrc |
| Announces truncationinvariant | holdssrc Returns next_offset. |
| Continuation hint | next_offsetsrc |
| Reports total lines | ? |
| Out-of-range offset | ? |
| Non-text | |
| Binary content never enters context silentlyinvariant | ? |
| Binary files | ? |
| Images | separate vision_analyzesrc |
| PDFs | ? |
| Notebooks | ? |
| Giant single line | char budget absorbs itsrc |
| Errors | |
| Missing path | ? |
| Path is a directory | ? |
| Empty file | ? |
| Non-vision model | ? |
| Correctness | |
| Mixed line endings | ? |
| Re-read dedup | (path,offset,limit)→mtime, 'File unchanged since last read'src |
| Staleness check before writeinvariant | holdssrc Compares mtime on write. |
| Line-number contamination guard | write-side detector rejects N| contentsrc |