01Safety
30.0 / 30
What changed in the harness
Selection accuracy 100→100, token cost up 2%, 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
29.1 / 30
03Economics
16.6 / 20
04Discoverability
16.8 / 20
Highest-impact fix
Estimated gain +3 pointsReduce the tool-definition and task token payload
Consolidate overlapping tools and trim repetitive schema text while preserving behavior, parameter, and return semantics.
Description evidence
Defects and rewrites.
2 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_skillselion |
no_return_description |
Search Skillselion, a curated directory of Claude Code agent skills, MCP servers and plugin marketplaces, ranked by real community signal (installs + GitHub stars). Use to browse trusted skills/MCPs/marketplaces for the task at hand (e.g. "postgres", "code review", "playwright"). To actually USE a skill, call load_skill (with an id from here, or directly with a query - it searches for you). Returns a ranked list of matching entries, each with its id, name, type, description, and community signal (installs/stars), for review or use as input to load_skill. |
top_skillselion |
no_return_description |
List the top entries on Skillselion (the leaderboard), optionally filtered to one listing type via `type`. Skills rank by installs; MCP servers and marketplaces rank by GitHub stars. Returns an ordered list (up to `limit`, default 10) of entries with id, name, type, and their ranking metric (install count or star count). |
Selection evidence
Confusable tool pairs.
2 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
load_skill |
synthesize_skills |
medium | For broad tasks like 'how should I approach testing in this repo', an agent could plausibly call load_skill (single best match) instead of synthesize_skills (cross-source digest), since both accept query+context and cover overlapping 'best practices' phrasing; descriptions do differentiate but require careful reading. |
search_skillselion |
top_skillselion |
low | A vague request like 'show me good MCP servers' could trigger either search (query-driven) or top (leaderboard, no query needed), but top_skillselion has no required query while search does, and their purposes (search vs ranked list) are distinct enough that confusion is unlikely except for very generic browsing requests. |
Compare the field