Tool tree
gradeNot graded
get_foundry_stateAnswer
Read the user-authored problem brief, active LaunchPad stage, record counts, quality warnings, and selected candidate. Call this first so every later action uses the human’s actual problem. This is a pure read and does not change the workspace.
View tool JSON
{
"name": "get_foundry_state",
"kind": "answer",
"impl": "imperative",
"description": "Read the user-authored problem brief, active LaunchPad stage, record counts, quality warnings, and selected candidate. Call this first so every later action uses the human’s actual problem. This is a pure read and does not change the workspace.",
"inputSchema": {
"type": "object",
"properties": {},
"additionalProperties": false
}
}research_and_ideateAction
Run the same server-side AI web-research workflow as the page: search the web for the actual problem, retain citation-linked supporting and counter evidence, build one problem-specific recommendation, stress-test it, and finalize the visible blueprint. Agent-started runs pause for visible, version-bound human approval before qualifying evidence and again before finalization. A request for a specific live result may pause for essential missing details instead of inventing an answer.
View tool JSON
{
"name": "research_and_ideate",
"kind": "act",
"impl": "imperative",
"description": "Run the same server-side AI web-research workflow as the page: search the web for the actual problem, retain citation-linked supporting and counter evidence, build one problem-specific recommendation, stress-test it, and finalize the visible blueprint. Agent-started runs pause for visible, version-bound human approval before qualifying evidence and again before finalization. A request for a specific live result may pause for essential missing details instead of inventing an answer.",
"inputSchema": {
"type": "object",
"properties": {
"problem_statement": {
"type": "string",
"description": "Problem to research. Omit when the user has already submitted it in the page."
}
},
"additionalProperties": false
}
}update_problem_briefAction
Create or revise the active structured problem brief. This visibly updates the Problem Hopper and moves the workspace to PROBLEM_DEFINED.
View tool JSON
{
"name": "update_problem_brief",
"kind": "act",
"impl": "imperative",
"description": "Create or revise the active structured problem brief. This visibly updates the Problem Hopper and moves the workspace to PROBLEM_DEFINED.",
"inputSchema": {
"type": "object",
"properties": {
"problem_type": {
"type": "string",
"description": "Kind of decision or opportunity."
},
"problem_statement": {
"type": "string",
"description": "Concrete problem statement. Required when the brief is empty."
},
"target_audience": {
"type": "string",
"description": "Specific affected audience."
},
"desired_outcome": {
"type": "string",
"description": "Observable outcome sought."
},
"current_behavior": {
"type": "string",
"description": "What the audience does today."
},
"constraints": {
"type": "array",
"description": "Implementation, policy, resource, or timing constraints.",
"items": {
"type": "string"
},
"uniqueItems": true
},
"geography": {
"type": "string",
"description": "Relevant geography."
},
"timeframe": {
"type": "string",
"description": "Decision or delivery timeframe."
},
"excluded_approaches": {
"type": "array",
"description": "Approaches that must not be recommended.",
"items": {
"type": "string"
},
"uniqueItems": true
},
"decision_criteria": {
"type": "array",
"description": "Criteria used to compare recommendations.",
"items": {
"type": "string"
},
"uniqueItems": true
}
},
"additionalProperties": false
}
}