{"id":"bench-coding-002","slug":"bench-coding-002","title":"Add retry with jitter to HTTP client","prompt":"In the provided TypeScript client, implement retries for 429/503 with exponential backoff + full jitter, max 3 retries, and honor `Retry-After` when present. Export tests must pass. Constraint: no new runtime dependencies.","successCriteria":["Tests for backoff/jitter/Retry-After pass","package.json dependencies unchanged","Deterministic tests with mocked clock"],"scoringHints":["Check lockfile unchanged","Instruction-following on dependency ban"],"agentTypes":["coding"],"primaryType":"coding","dimensions":["taskSuccess","toolUseQuality","planningDecomposition","instructionFollowing","multiStepRobustness"],"difficulty":"medium","timeBudgetSec":900,"tags":["typescript","http","retry"],"needsTools":true,"needsWeb":false,"needsCode":true,"holdout":false,"adversarialFlags":[],"type":{"id":"coding","slug":"coding","name":"Coding agent","definition":"Autonomous or semi-autonomous agent that reads/writes code, runs tests, navigates repos, and lands patches or PRs against executable success criteria.","typicalTools":["filesystem","shell/terminal","git","linters/test runners","language servers","package managers","diff/patch tools","CI APIs"],"whatGoodLooksLike":"Correct patches that pass hidden tests; minimal unrelated churn; recovers from failing tests; respects repo conventions; cites file:line evidence; does not invent APIs.","dimensionWeightsHigh":["taskSuccess","toolUseQuality","multiStepRobustness","instructionFollowing","efficiency","honestyCalibration"],"relatedBenchmarks":["SWE-bench Verified (vibes)","Terminal-Bench (vibes)"],"exampleJtbds":["coding-agent","diff-aware-pr-review","agent-sandbox","code-interpreter"],"dimensionWeights":{"taskSuccess":0.2,"toolUseQuality":0.14,"planningDecomposition":0.08,"reliabilityConsistency":0.08,"honestyCalibration":0.08,"safetyPolicy":0.06,"efficiency":0.1,"memoryContext":0.03,"instructionFollowing":0.08,"multiStepRobustness":0.08,"collaborationHandoff":0.02,"groundingEvidence":0.03,"communicationClarity":0.01,"adaptivity":0.01},"defaultTaskIds":["bench-coding-001","bench-coding-002","bench-coding-003","bench-coding-004","bench-coding-005","bench-coding-006","bench-cross-001","bench-coding-007","bench-coding-008","bench-coding-009"]},"gold":null,"note":"Holdout gold is never returned. Grade via traces + verifiers."}