Public leaderboard

Public assessment

ofershap/mcp-server-github-gist (mcp-server-github-gist)

mcp-server-github-gist · v0.1.0 · scanned

What changed in the harness

Selection accuracy 95→95, token cost up 24%, 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

24.1 / 30

24.1 out of 30
03Economics

20.0 / 20

20.0 out of 20
04Discoverability

12.1 / 20

12.1 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.

8 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
gist_list
name_restates_behavior params_unexplained no_return_description
Fetch a paginated list of Gists owned by the authenticated user, sorted by most recently updated. 'page' selects which page of results to retrieve (starting at 1); 'perPage' controls how many gists are returned per page. Returns an array of gist summaries including id, description, public/private status, and file names.
gist_get
name_restates_behavior params_unexplained no_return_description
Retrieve full details of a single Gist, including the contents of every file it contains. 'gistId' is the unique identifier of the Gist to fetch. Returns the gist's metadata (description, visibility, owner, timestamps) plus each file's name and content.
gist_create
name_restates_behavior no_return_description
Create a new Gist owned by the authenticated user from the given files. 'files' maps filenames to their content and is required; 'description' sets an optional summary shown on the gist; 'public' controls whether the gist is publicly listed (defaults to false, i.e. secret). Returns the newly created gist's id, URL, and file listing.
gist_update
name_restates_behavior no_return_description
Modify an existing Gist's description and/or file contents. 'gistId' identifies the gist to update; 'description' replaces its current description if provided; 'files' maps filenames to new content, where setting a filename's value to null deletes that file. Returns the updated gist's metadata and current file listing.
gist_delete
name_restates_behavior no_return_description
Permanently delete a Gist owned by the authenticated user. 'gistId' identifies the gist to delete. Returns confirmation that the gist was deleted; this action cannot be undone.
gist_starred
params_unexplained no_return_description
List Gists the authenticated user has starred, paginated. 'page' selects which page of results to retrieve (starting at 1); 'perPage' controls how many gists are returned per page. Returns an array of starred gist summaries including id, description, owner, and file names.
gist_star
name_restates_behavior no_return_description
Add a star to a Gist on behalf of the authenticated user, marking it as favorited. 'gistId' identifies the gist to star. Returns confirmation that the gist was starred.
gist_unstar
name_restates_behavior no_return_description
Remove the authenticated user's star from a Gist. 'gistId' identifies the gist to unstar. Returns confirmation that the gist was unstarred.

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
gist_list gist_starred medium A vague request like 'show me my gists' or 'list my favorite gists' could be routed to either the general gist list or the starred-only list if the agent doesn't parse whether the user means all gists or specifically starred ones.
gist_delete gist_unstar medium A task phrased as 'remove this gist' or 'get rid of gist X' is ambiguous between permanently deleting the gist and merely unstarring/removing it from a starred list, since both are 'removal' actions on a gistId.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard