Agent · Nous Research · framework
Hermes
github.com/NousResearch/hermes-agent
21
tools registered
6/6
compared capabilities
it has a tool for
it has a tool for
1
of those tools
read from source
read from source
Hermes registers 21 tools, covering 6 of 6 compared capabilities.
Surface size
tools registered, across the 13 harnesses hereSurface size, not quality. Every registered tool spends context on every request, so a longer bar is a cost as much as a capability — and a harness that reaches a capability through a general tool may register fewer.
Capabilities
what serves each, and where it is compared| Capability | Tools | Across harnesses |
|---|---|---|
| Read | read_file | 13 of 13 have one → |
| Browser | browser_navigate | 3 of 13 have one → |
| Web search | web_search | 9 of 13 have one → |
| Code search | search_files | 13 of 13 have one → |
| Shell | processpartial · terminal | 13 of 13 have one → |
| Long-term memory | memory | 4 of 13 have one → |
Where its tools differ from the rest
- read_file · Staleness check before writeinvariantholdsvs fails in 6 of 8
- read_file · Byte / char cap100k chars (not lines)vs 50 KB in 3 of 5
- read_file · Whichever hits firstcharsvs lines or bytes in 4 of 5
- read_file · 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.
All 21 tools
names verbatim, as registeredNot yet compared means registered by Hermes but with no comparison page here — todo lists, subagent spawning, approval flows. They become capabilities the day there is a comparison worth reading, and the data does not have to change for that to happen.
Grok Build
Kimi Code →