test-corpus

An index of PDF-centric corpora

A Qiqqa Test Library / Test Corpus which contains various PDF document samples, etc. collected from live Qiqqa libraries to showcase issues and check regressions in the software.

Collection of certificates and keys with different encodings and encryptions

A synthetic codebase built to be hard to index — 800 files, 72 file types, comments in 8 writing systems. Measure a context tool against something difficult without pointing it at your own code. MIT.

the gauntlet corpus

An open, reproducible PDF corpus for benchmarking and regression testing across real-world, synthetic, malformed, accessibility, archival, rendering, parsing, repair, and conformance cases.

CSV parser conformance test corpus

A false-positive corpus for secret scanners: 57 formats of ordinary log and build output, zero credentials. Every secret your scanner reports here is a false positive.

An adversarial conformance corpus for IFC parsers — 52 trap files, and a scoreboard of how IfcOpenShell, web-ifc and friends survive them.

Shared test corpus and harness for YINI parser implementations — validating clear, indentation-insensitive configuration with nested sections and predictable parsing.

Language-agnostic JSON test vectors for NIP behavioral interpretation

539 regex constructs run on 16 real engines. 264 behave differently somewhere and only 83 tell you. MIT-licensed, reproducible.

Shared StructuredMerge conformance fixtures for peer language implementations

A repo of 50+ languages to test the detection capabilities of repo-scanners. Built to stress-test any multi-language scanner, prove regression resistance, and validate structural signature extraction at scale.

Reproducible benchmark and test corpus for prompt-injection defenses — adapters for common sanitization libraries, scored leaderboard, CI-ready. Security-adjacent, evergreen.