| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-08-06 | 1.1 kB | |
| v3.6.8 -- Ollama fix, empty-evidence validation, single-model warning source code.tar.gz | 2026-08-06 | 421.3 kB | |
| v3.6.8 -- Ollama fix, empty-evidence validation, single-model warning source code.zip | 2026-08-06 | 750.9 kB | |
| Totals: 3 Items | 1.2 MB | 0 | |
v3.6.8 — Bugfix: local model robustness + validation pipeline
- Better Ollama/local provider error messages. Connection-refused and timeout errors now name the provider, URL, and suggest checking if the server is running.
- Empty-evidence findings skip the vote and go straight to
needs-review. Findings with no evidence bypass the adversarial validator (which always rejects per its prompt) and are flagged for human review instead of being silently dropped. - Single-model + vote_n=1 warning. Emits a warning that validation is weaker when same model validates its own findings.
- JSON parse resilience for local models.
extract_findingsnow logs when a model returns text but no parseable JSON (previously silent drop — 0 findings with no diagnostic). Also auto-fixes trailing-comma JSON ([...,]) which small models commonly produce. - Visible diagnostics when agents return 0 findings. Pipeline emits the response tail so the operator can see what the model actually returned.
Full changelog: https://github.com/JoasASantos/NeuroSploit/compare/v3.6.7...v3.6.8
