01Safety
30.0 / 30
What changed in the harness
Selection accuracy 100→100, token cost up 5%, unconfirmed writes 100%→100%.
Category breakdown
Where the score comes from.
Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.
02Legibility
29.0 / 30
03Economics
20.0 / 20
04Discoverability
20.0 / 20
Highest-impact fix
Estimated gain +1 pointRewrite 3 underspecified tool descriptions
State each tool's behavior, explain every parameter, and declare return semantics, prioritizing tools exercised by failed selection tasks.
Description evidence
Defects and rewrites.
3 defects found across the exposed tool descriptions. Suggested rewrites make purpose, inputs, boundaries, and returns easier for an agent to understand.
| Tool | Defect types | Suggested rewrite |
|---|---|---|
search_mcp_servers |
params_unexplained |
Search Remote OpenClaw's directory of 13,870+ MCP servers for Claude Code, OpenClaw, Hermes, Codex, and Cursor. Returns names, descriptions, links, and popularity. Pass a query such as 'postgres database' and an optional limit for the maximum number of results to return. |
search_skills |
params_unexplained |
Search Remote OpenClaw's directory of 4,384+ agent skills for Claude Code, OpenClaw, Hermes, and Codex. Returns names, descriptions, links, and install commands. Pass a query such as 'pdf' or 'react' and an optional limit for the maximum number of results to return. |
search_plugins |
params_unexplained |
Search Remote OpenClaw's directory of Claude Code / OpenClaw / Hermes / Codex plugins. Returns names, descriptions, and links. Pass a query such as 'testing' and an optional limit for the maximum number of results to return. |
Selection evidence
Confusable tool pairs.
1 pair where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
search_skills |
search_plugins |
medium | Both directories target the same platforms (Claude Code, OpenClaw, Hermes, Codex) and a task like 'find a pdf skill' or 'I need a testing skill' could plausibly match either, since users frequently conflate agent skills with plugins and neither query term ('skill' vs 'plugin') is strongly disambiguating in natural language. |
Compare the field