DecisionLens
An offline, zero-dependency AI decision agent built on the Python standard library. It accepts natural-language questions and analyzes CSV data, attaching traceable evidence references—including file, line number, and field name—to each inference. Reasoning is performed using deterministic algorithms rather than large language models.
Timeline (last 90 days)
DecisionLens is an offline, zero-dependency decision agent written entirely in the Python standard library.
DecisionLens attaches evidence citations to every assertion, including source file, line number, and field name.
DecisionLens runs a classic perceive→reason→explain loop consisting of six steps: PERCEIVE, INDEX, RETRIEVE, REASON, EXPLAIN, and ACT.
DecisionLens completely bypasses language models, replacing probabilistic reasoning with deterministic algorithms where every computational step is an explicit arithmetic operation.
DecisionLens's test suite contains 8 assertions, all of which pass (8/8) when running python eval.py.
DecisionLens's test suite asserts that the code makes no network calls and can run on an air-gapped machine.
DecisionLens uses only the Python standard library (csv, argparse, datetime), requires no pip installation, no API key, and has no cloud dependencies.
DecisionLens has fewer than 1,000 lines of code and supports HTML report output.
The DecisionLens project repository is at github.com/Zhiyilang/decisionlens and comes with three sample datasets and a complete evaluation suite.
DecisionLens was created as an entry for the Open Agent Hackathon 2026 (Track 02: Explainable Multi-Source Reasoning).
1 more timeline events
All Facts (11)
Making non-hallucination into an automatically runnable assertion means developers can verify that this safety property has not been broken every time they modify the code.
50%UnverifiedDecisionLens attaches evidence citations to every assertion, including source file, line number, and field name.
50%UnverifiedDecisionLens runs a classic perceive→reason→explain loop consisting of six steps: PERCEIVE, INDEX, RETRIEVE, REASON, EXPLAIN, and ACT.
50%UnverifiedDecisionLens completely bypasses language models, replacing probabilistic reasoning with deterministic algorithms where every computational step is an explicit arithmetic operation.
50%UnverifiedDecisionLens's test suite contains 8 assertions, all of which pass (8/8) when running python eval.py.
50%UnverifiedDecisionLens's test suite asserts that the code makes no network calls and can run on an air-gapped machine.
50%UnverifiedDecisionLens uses only the Python standard library (csv, argparse, datetime), requires no pip installation, no API key, and has no cloud dependencies.
50%UnverifiedDecisionLens has fewer than 1,000 lines of code and supports HTML report output.
50%UnverifiedThe DecisionLens project repository is at github.com/Zhiyilang/decisionlens and comes with three sample datasets and a complete evaluation suite.
50%UnverifiedDecisionLens is an offline, zero-dependency decision agent written entirely in the Python standard library.
50%UnverifiedDecisionLens was created as an entry for the Open Agent Hackathon 2026 (Track 02: Explainable Multi-Source Reasoning).
50%