Repository Issue Activity (beta)

#165 sn-image-base: Consider env-configurable image model instead of hardcoded default

#160 export_pptx guards silently skip validation because of file-name mismatch (`style_spec.json` vs `style-spec.json`)

#150 security issues by NVIDIA spector scanned result.

#145 `caption_images.py` crashes with `JSONDecodeError` when `raw_documents.json` is corrupted

#136 Docs idea: optional first run for sn-infographic

#128 parse_user_docs.py: dead loop iterating tables for no purpose

#127 `lstrip("PMCpmc")` in pubmed_search.py strips characters, not a prefix

#126 `cp -r` install instructions silently overwrite existing skills

#121 sn_agent_runner.py: `--image-size` accepts only `"2k"` but is declared a `choices=` of length 1

#122 caption.py uses MD5 for cache keys — fine, but worth a one-line comment

#120 Hey — small thing I noticed while reading through the search scripts.

#97 Design question: is the `sn_agent_runner` Level 2 discovery intentionally omitted everywhere except the comments?

#107 `_STYLE_DIMENSIONS_PATH` points to a directory that doesn't exist — repair logic is dead

#104 bugs? search scripts just catch `httpx.HTTPStatusError`

#94 arXiv search query doesn't quote multi-word terms — spaces break field-qualified queries

#92 `caption_batch()` has no concurrency — `caption_images.py` solved this, standalone script didn't

#93 Clean, generic `_resolve_prompt` helper — happy to see this pattern

#91 🐛 nano_banana: `None` reaches `base64.b64decode` when response is missing inlineData fields