Build a Gas Pipeline Tender Risk Agent with OpenAI Agents SDK & Galileo
Pipeline project bid delays create cost overruns due to complex procurement signals across North Africa. Build an autonomous agent using OpenAI Agents SDK and Galileo to evaluate tender delay risk and calculate project risk scores exceeding 90% accuracy.
What you are building
The core problem, expected build, and operating context for this challenge.
Utilize OpenAI Agents SDK with Galileo evaluation tracing to analyze cross-border pipeline procurement documents and generate risk mitigation schedules.
How work is evaluated
Evaluates the agent's ability to extract bid delay parameters and output formatted risk scores verified by Galileo metrics.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
test_risk_score_range
Ensure calculated risk score falls between 0 and 1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
extraction_precision
Precision of extracted tender parameters • target: 0.95 • range: 0.85-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Parse unstructured gas pipeline tender addendums and bilingual bid extensions
Implement autonomous risk scoring agents using OpenAI Agents SDK tool invocation
Evaluate prompt accuracy and response hallucinations using Galileo observability metrics
Produce automated schedule mitigation recommendations for energy megaprojects
Reference links and supporting material
Synthetic dataset of Mediterranean gas pipeline tender notices, extension addendums, and historical concession agreement risks.
Challenge setup
Evaluates the agent's ability to extract bid delay parameters and output formatted risk scores verified by Galileo metrics.
Challenge input
JSON containing tender text and historical delay data
OpenAI Agents SDK
Official framework for multi-turn structured task execution
Galileo
Provides prompt and hallucination metrics
Evaluated output
JSON with delay_days, risk_score, and mitigation_steps
- Ensure calculated risk score falls between 0 and 1
- Precision of extracted tender parameters • target: 0.95 • range: 0.85-1
- Benchmark: ProcurementRiskBench-GCC
- Extraction Precision target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- OpenAI · Optional
- Bito AI · Optional
- Galileo · Required
- Galileo · Required
Evaluation contract
- test_risk_score_range · Weight 1
- extraction_precision · Weight 1
Metrics
- extraction_precision · Target 0.95
Benchmarks
- ProcurementRiskBench-GCC
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing tender text and historical delay data
Output: JSON with delay_days, risk_score, and mitigation_steps
Execution harness: Python
Evaluation tools
- OpenAI Agents SDK · Agent Framework
- Galileo · Evaluation Engine
Recipe freeze marker
No recipe freeze time is recorded.