01Safety
0.0 / 30
What changed in the harness
Selection accuracy 98→93, token cost up 5%, 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.
02Legibility
26.8 / 30
03Economics
18.8 / 20
04Discoverability
13.7 / 20
Highest-impact fix
Estimated gain +30 pointsAdd 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 |
|---|---|---|
ask |
no_return_description |
Executes GitHub Copilot CLI with a given prompt, optional file/directory context, tool and MCP permission controls, and session/resume options. Returns Copilot CLI's textual response along with execution metadata such as logs and exit status. |
batch |
no_return_description |
Delegates an array of atomic tasks (each with a description, optional target files/dirs, and priority) to GitHub Copilot CLI for sequential or parallel batch processing. Returns per-task results including success/failure status and any output or error produced by Copilot CLI for each task. |
review |
no_return_description |
Runs GitHub Copilot CLI against target files/directories to perform a code review of a specified type (security, performance, quality, etc.), with configurable severity, output format, and issue limits. Returns a review report listing found issues, severities, and optional fix suggestions and priority rankings in the requested output format. |
ping |
no_return_description |
Echoes back the provided prompt message, useful for verifying the MCP server connection is alive. Returns the same text that was passed in the prompt parameter. |
Help |
name_restates_behavior no_return_description |
Retrieves the GitHub Copilot CLI's built-in help text, listing available commands, flags, and usage guidance. Returns that help text as a string. |
version |
name_restates_behavior no_return_description |
Retrieves the installed Copilot CLI version and MCP server version along with related system/environment details. Returns this information as a formatted text block. |
brainstorm |
no_return_description |
Generates creative ideas for a given prompt using a selectable brainstorming framework (divergent, convergent, scamper, design-thinking, lateral, or auto), optionally scoped by domain, constraints, and existing context. Returns a list of generated ideas, optionally with feasibility/impact analysis for each. |
timeout-test |
name_restates_behavior no_return_description |
Blocks for the specified duration (in milliseconds, minimum 10ms) to verify that long-running MCP calls do not get prematurely timed out. Returns a confirmation once the specified duration has elapsed, including the actual elapsed time. |
Selection evidence
Confusable tool pairs.
0 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
| No confusable tool pairs were flagged in this assessment. | |||
Compare the field