Sunglasses — AI skill for Claude Code

SUNGLASSES

[](https://github.com/sunglasses-dev/sunglasses/actions/workflows/pattern-integrity.yml) [](https://pypi.org/project/sunglasses/) [](https://github.com/sunglasses-dev/sunglasses/blob/main/.github/workflows/pattern-integrity.yml) [](LICENSE) [](https://pypistats.org/packages/sunglasses)

**Open source input firewall for AI agents, beta.** A local scanner checks text, code, PDFs, images, QR codes, audio and video with 1,540 patterns across 118 categories and reports findings and incomplete scans. A Claude Code hook blocks credential leaks and policy violations before tools run.

Sunglasses is a local, open-source scanner for text and supported files. It reports what it matched **and what it could not read**, so you can decide what to pass onward. It produces findings and an exit status; a CI job, a Claude Code hook or your own code acts on that result.

Four exit statuses, deliberately different signals:

exit	meaning
0	inspection completed in the supported scope, and nothing matched
1	threat found — the inspection may still have been incomplete, and that is reported alongside
3	incomplete — nothing matched in the part that was inspected; some component was not
2	usage or operational error

**`0` and `3` never collapse into each other.** "Everything I support reading here was read, and nothing matched" and "this format was not inspected" are different facts, and the second one is where agents get hurt. Exit `0` is not a guaran