Public leaderboard

Public assessment

atlasprzetargow/mcp-server (@atlasprzetargow/mcp)

atlasprzetargow-mcp · v0.1.0 · scanned

What changed in the harness

Selection accuracy 100→100, token cost up 3%, unconfirmed writes 0%→0%.

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.

01Safety

0.0 / 30

0.0 out of 30
02Legibility

28.8 / 30

28.8 out of 30
03Economics

19.6 / 20

19.6 out of 20
04Discoverability

13.8 / 20

13.8 out of 20

Highest-impact fix

Estimated gain +30 points

Add explicit identity and permission preflight tools

Expose machine-readable principal/tenant confirmation and a non-mutating permission check so agents can verify both before destructive actions.

Description evidence

Defects and rewrites.

1 defect 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
get_category_stats
no_return_description
Retrieve aggregate statistics for a CPV category based on historical BZP+TED data. Returns the tender count, average and median value, average number of offers, and average deadline period for the given CPV code, over the selected window. Useful for market sizing and competitive benchmarking.

Selection evidence

Confusable tool pairs.

3 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.

Tool A Tool B Confidence Why they collide
get_buyer get_contractor high Both accept a 10-digit NIP and return an entity profile with statistics and winning counterparties; the only difference is the entity role (zamawiający vs wykonawca). A task like 'pobierz profil firmy o NIP 526...' or 'z kim ta firma wygrywa przetargi' doesn't specify the role, and many companies act as both buyers and contractors, making the wrong choice plausible.
get_category_stats get_province_stats medium Both are aggregate-statistics tools over BZP+TED data, and a phrase like 'statystyki przetargów budowlanych' or 'statystyki rynku IT w Polsce' could mean statistics dimensioned by CPV category or by province; get_province_stats even returns top CPV categories per city, blurring the dimension boundary.
search_tenders search_entities medium Both are full-text search tools returning result lists, so a query naming a company or buyer (e.g. 'wyszukaj GDDKiA', 'co w przetargach robi Budimex') is ambiguous between finding the entity's profile/NIP and finding the tenders it publishes or bids on, especially since search_tenders also supports buyerNip filtering.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard