Awesome-list pitch kit
Awesome-list pitch kit
Status: drafts ready — say “send the pitches” and PRs go out from the repo’s GitHub account.
1. Jenqyang/Awesome-AI-Agents → Benchmark/Evaluator section
URL: https://github.com/Jenqyang/Awesome-AI-Agents
PR title: Add STRATA-Bench to Benchmark/Evaluator
PR body:
Adds STRATA-Bench, a 36-task open-source benchmark (Apache-2.0) for AI agents on fragmented spatial-temporal market intelligence. It scores whether agents disclose, tag, refuse, or model uncertainty across six failure families (geographic masking, temporal attrition, OCR laundering, silent interpolation, non-uniform axes, source blending) instead of hallucinating downtown figures or plotting gaps as data. Ships a self-contained synthetic corpus, held-out gold, three reference agents, and a community leaderboard with auto-scoring.
Entry line (insert in Benchmark/Evaluator list):
- [STRATA-Bench](https://github.com/movahedi-ca/strata-bench) - Benchmark for AI agents on fragmented spatial-temporal market intelligence; scores whether agents disclose, tag, refuse, or model instead of hallucinating. []
2. reacher-z/awesome-agent-observability → Benchmarks section
URL: https://github.com/reacher-z/awesome-agent-observability
PR title: Add STRATA-Bench to Benchmarks
PR body:
Adds STRATA-Bench (Apache-2.0): 36 sandbox tasks scoring whether AI agents disclose, tag, refuse, or model uncertainty when compiling longitudinal market intelligence from fragmented, geographically nested, temporally incomplete evidence. Hard fails (geographic masking, silent interpolation, paywall fabrication) cap tasks at 55%. Self-contained synthetic corpus, three reference agents, community leaderboard with bot auto-scoring.
Entry line:
- [STRATA-Bench](https://github.com/movahedi-ca/strata-bench) — 36-task benchmark scoring whether AI agents disclose, tag, refuse, or model uncertainty on fragmented spatial-temporal market intelligence.
3. ARUNAGIRINATHAN-K/awesome-ai-agents-2026 → Agent Evaluation and Benchmarks
URL: https://github.com/ARUNAGIRINATHAN-K/awesome-ai-agents-2026
PR title: Add STRATA-Bench to Agent Evaluation and Benchmarks
PR body:
Adds STRATA-Bench: 36-task Python benchmark (Apache-2.0) evaluating whether AI agents can compile longitudinal market intelligence from fragmented spatial-temporal evidence without relabeling metro figures as downtown, silently interpolating missing quarters, or plotting gaps as data. Includes held-out gold, reference agents, and an auto-scored community leaderboard.
Entry line:
- STRATA-Bench `🌱` `[Python]` `[Benchmark]` - 36-task benchmark for AI agents on fragmented spatial-temporal market intelligence; scores honesty about gaps.