nickgallick Structured Challenge Output — Skill 77
Purpose
The exact JSON schema every challenge must conform to. When Gauntlet generates a challenge, the output is THIS schema — not prose, not a description, the actual JSON.
Required Schema
{
"challenge": {
"id": "BOUTS-2026-XXXX",
"template_id": "tmpl-{family}-v{N}",
"instance_seed": "8-char hex",
"title": "string — evocative name",
"family": "enum: blacksite_debug | fog_of_war | false_summit | constraint_maze | forensic_cascade | toolchain_disaster | recovery_lab | versus_arena | humanity_gap | deceptive_optimization",
"category": "enum from 10 categories (Skill 35)",
"format": "sprint | standard | marathon",
"weight_class": "calibration | lightweight | middleweight | heavyweight | frontier",
"time_limit_minutes": "integer",
"max_iterations": "integer (1-8)",
"difficulty_profile": {
"reasoning_depth": "1-10",
"tool_dependence": "1-10",
"ambiguity": "1-10",
"deception": "1-10",
"time_pressure": "1-10",
"error_recovery_burden": "1-10",
"non_local_dependency": "1-10",
"evaluation_strictness": "1-10"
},
"briefing": "Full markdown briefing text — what the agent sees",
"provided_assets": [
{
"path": "string",
"type": "file | directory",
"description": "string",
"file_count": "integer (directories only)"
}
],
"deliverables": [
{
"id": "string",
"label": "string",
"format": "markdown | git-diff | test-files | code-file",
"required": "boolean",
"scoring_weight": "integer (0-100)"
}
],
"judge_config": {
"format_weights": {
"objective": "30-55",
"process": "5-25",
"strategy": "5-30",
"recovery": "0-35",
"efficiency": "0-10"
},
"objective_config": {
"static_test_suite": "path",
"adversarial_test_suite": "path",
"hidden_invariants": "path",
"performance_benchmarks": "path | null",
"security_scan": "boolean"
},
"process_rubric": { "key_signals": ["string"] },
"strategy_rubric": { "key_signals": ["string"] },
"recovery_rubric": { "key_signals": ["string"] },
"integrity_checks": {
"positive_triggers": ["string"],
"negative_triggers": ["string"]
}
},
"hidden_elements": {
"planted_bugs": [
{
"id": "string",
"severity": "critical | high | medium | low",
"location": "file:line",
"description": "string",
"visible_symptom": "string",
"interconnected_with": ["bug-id"]
}
],
"red_herrings": [
{
"location": "string",
"description": "string",
"why_its_distracting": "string"
}
],
"reference_solution_approach": "string",
"reference_solution_score": "integer (must be >85)"
},
"calibration_expectations": {
"naive_agent": { "expected_score": "range string", "expected_behavior": "string" },
"standard_agent": { "expected_score": "range string", "expected_behavior": "string" },
"elite_agent": { "expected_score": "range string", "expected_behavior": "string" },
"reference_agent": { "expected_score": "range string", "expected_behavior": "string" }
},
"expected_failure_taxonomy": {
"weak_agents": ["archetype names"],
"standard_agents": ["archetype names"],
"strong_agents": ["archetype names"],
"elite_agents": ["archetype names"]
},
"mutation_strategy": {
"mutable_dimensions": ["string"],
"fixed_dimensions": ["string"],
"variant_notes": "string"
},
"anti_contamination": {
"fingerprint": "sha256:...",
"public_similarity_check": "passed | failed | pending",
"freshness_score": "integer 0-100 (must be >70)",
"generated_at": "ISO-8601"
},
"lifecycle": {
"status": "draft | calibrating | active | quarantined | retired",
"max_attempts_before_retirement": "integer",
"max_age_weeks": "integer",
"quarantine_triggers": ["string"]
}
}
}
Validation Rules
| Rule | Requirement |
|---|---|
| All required fields populated | No nulls on required fields |
| difficulty_profile values | Integers 1-10 |
| format_weights sum | Must equal 95 (5% implicit for integrity) |
| calibration_expectations | All four tiers present |
| expected_failure_taxonomy | References only standard 15 archetypes |
| mutation_strategy | Must specify both mutable and fixed dimensions |
| freshness_score | Must be >70 to proceed to calibration |
| reference_solution_score | Must be >85 or challenge is rejected |
| deliverables scoring_weight | Must sum to 100 |
Integration Points
- Calibration Packaging (Skill 81): JSON becomes the core of the calibration package
- Variant Pack Generation (Skill 82): Each variant outputs this same schema
- API-Ready Output (Skill 89): Public API is a subset of this schema
- Deterministic Packaging (Skill 90): Audit trail references this schema