01Safety
0.0 / 30
What changed in the harness
Selection accuracy 98→100, token cost up 13%, 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
27.1 / 30
03Economics
20.0 / 20
04Discoverability
13.2 / 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.
6 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 |
|---|---|---|
megalaunch_status |
no_return_description |
Check MegaLaunch service status and recent launch statistics. Returns the current service health/availability and stats on recent launches (e.g., counts and outcomes). |
megalaunch_pricing |
no_return_description |
Get current pricing for token launch packages (Basic and Premium). Returns the price and included features for each package. |
megalaunch_create_order |
params_unexplained |
Create a new meme token launch order on Solana/pump.fun. Params: package (basic or premium tier), tokenName (custom token name), tokenSymbol (1-10 character ticker), tokenDescription (text shown as the token's public description/blurb), aiPick (let AI choose a trending name/symbol instead of custom values), theme (theme guiding AI-generated token identity when aiPick is used). Returns the deposit address and amount to send to fund the order. |
megalaunch_list_orders |
no_return_description |
List your token launch orders with optional status filter, limit, and offset for pagination. Returns an array of matching orders with their IDs, statuses, and summary details. |
megalaunch_get_order |
no_return_description |
Get detailed status of a specific token launch order by orderId. Returns the order's full details, including current status, package, token info, and payment/deployment progress. |
megalaunch_cancel_order |
no_return_description |
Cancel a pending token launch order by orderId (only works for unpaid orders). Returns the updated order confirming the cancelled status. |
Selection evidence
Confusable tool pairs.
4 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
megalaunch_list_orders |
megalaunch_get_order |
high | A vague request like 'check the status of my order' or 'show my order' doesn't specify whether to fetch one order by ID or list all orders (optionally filtered by status); both tools surface order status and share the 'status' concept, so an agent lacking an explicit orderId may wrongly call list_orders instead of get_order or vice versa. |
megalaunch_status |
megalaunch_list_orders |
medium | A task phrased as 'what's the status of my launch(es)' is ambiguous between megalaunch_status (service-wide stats) and megalaunch_list_orders (the user's own orders with a status filter) — both surface 'status' info at different scopes, and the wording doesn't clearly indicate which the user wants. |
megalaunch_get_order |
megalaunch_cancel_order |
low | Both operate on a specific orderId and are the only two single-order-targeting tools, so a terse instruction like 'stop/close order X' could be misrouted to get_order instead of cancel_order if the agent misreads intent, though the verbs are usually distinguishable. |
megalaunch_create_order |
megalaunch_cancel_order |
low | Both are order-lifecycle mutating actions sharing the 'order' vocabulary; an ambiguous phrase like 'reset my launch order' could theoretically be misrouted, though 'create' vs 'cancel' intents are normally clear enough to avoid real confusion. |
Compare the field