33487964548
84%

Ran 01 Sep 2026 08:41AM UTC

Jobs 1

Files 43

Run time 1min

Badge

Embed ▾

Committed 01 Sep 2026 08:35AM UTC coverage: 84.045% (+0.02%) from 84.026%

Build # 33487964548

Build Type

push

github

Committed by

umputun

Commit Message

docs: fix short-message LLM and ban-path claims

the openai section claimed short messages are always checked by OpenAI
regardless of --openai.veto. An earlier spam verdict from stop-words, a meta
check, a lua plugin or CAS returns at detector.go:355 before any provider runs,
so an LLM is reached only when nothing has flagged the message. It also names
--gemini.check-short-messages now, which gates the same path.

the documents-only section attributed the permanent ban to CAS. Any of the
pre-gate checks can produce a spam verdict, and any spam verdict returns
PermanentBanDuration (app/bot/spam.go:172); CAS is only the one enabled on a
stock install. The same paragraph implied check-short-messages could bring an
LLM in for a caption-less file, which DocumentsCheck has already decided.

CLAUDE.md carried the same claim, and now states the non-short cases where an
enabled veto-mode provider does get to override: an appended parent bringing
req.Msg to at least MinMsgLen, or MinMsgLen set to 0.

the AdmitDocuments field comment named only the detector effect, omitting the
locator write that feeds --max-short-msg-count.

Related to #447

Coverage Stats

9134 of 10868 relevant lines covered (84.04%)

234.92 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
1	33487964548.1	01 Sep 2026 08:41AM UTC	43	84.04	GitHub Action Run

umputun / tg-spam / 33487964548
84%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 33487964548

umputun / tg-spam / 33487964548 84%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 33487964548

umputun / tg-spam / 33487964548
84%

README BADGES
x