Reddit AutoModerator Falsely Removing Positive Comments – Need Help Refining Comment Filter Rule

[GorillaWolf2099]

Hi there,

I'm configuring AutoModerator for a subreddit and running into an issue where it falsely removes well-intentioned comments—especially those praising fan cast ideas.

Example of a wrongly removed comment:

“I didn’t like scary movies, but I’d watch this person’s version of ANOES with the cast they’d chosen.”

This was flagged and removed, but there’s no obvious slur or insult in it. I suspect the current body (includes) rule might be overreaching or colliding with vague word matches.

---

Here’s a portion of the current AutoModerator config:

Remove Slurs

---

type: submission
title+body (regex):

• "zion(ist|ism)"
• "fag(gy|s)"
• "k[iy]ke"
• "be[ea]ner"
• "spick?"
• "honk(e|y)"
• "crack(a|er)"
• "wet[ -]?back"
• "gook"
• "wop(p)"
• "chink(y|ie)"
• "fagg?([eio]t)?(ry)?"
• "d[iy]ke"
• "trann(y|ie)"
• "n[ie]g(ger|ga|g?let|g?ro|g?uh?)"
• "porch ?monke?y"
• "ha[jg]i"
• "(towel|rag)[ -]?head"
• "jigg?[ae]r?boo"
• "kr[ao]ut(s)?"
  author:
  is_moderator: false
  action: report
  report_reason: "Slur keywords: {{match}}"

Remove Offensive Comments (Slurs, Politics, Misogyny, Inappropriate Content)

---

type: comment
body (includes):

Existing slurs

• "dumbass"
  • "zionist"
  • "fag"
  • "kike"
  • "beaner"
  • "spic"
  • "honky"
  • "cracker"
  • "wetback"
  • "gook"
  • "wop"
  • "chink"
  • "dyke"
  • "tranny"
  • "nigger"
  • "porch monkey"
  • "haji"
  • "towelhead"
  • "jigaboo"
  • "kraut"

  Political offensive words

  • "woke agenda"
  • "sjw"
  • "snowflake"
  • "libtard"
  • "feminazi"
  • "magat"
  • "leftard"
  • "commie scum"
  • "blue haired freak"
  • "soy boy"
  • "dei castings suck"

  Misogynistic & inappropriate phrases

  • "my wife she certainly knows how to handle a stick"
  • "wish she was my wife"
  • "should be my wife"
  • "get back to the kitchen"
  • "women belong in the kitchen"
  • "not like other girls"
  • "just a dumb blonde"
  • "she deserves a real man"
  • "she needs a real man"
  • "she’d look better without clothes"
  • "she’d look better without makeup"
  • "smash or pass"
  • "bet she’s good in bed"
  • "she's asking for it"
  • "daddy's little girl"
  • "she needs to be tamed"
  • "females these days"

  Inappropriate emoji shortcodes

  • "\U0001F924" # (drooling face)
  • "\U0001F975" # (hot face)
  • "\U0001F60D" # (heart eyes)
  • "\U0001F346" # (eggplant)
  • "\U0001F351" # (peach)
    author:
    is_moderator: false
    action: remove
    comment: |
    "This comment was removed for violating our community guidelines on offensive language and inappropriate remarks."
    report_reason: "Offensive content detected."

---

What I’ve Tried:

Reviewed mod logs, but unclear which keyword is being triggered.

Suspect overly broad matches or hidden word conflicts.

Considered switching to regex with word boundaries (\b) or using ~body exclusions.

Debating changing action: remove to filter for review instead.

---

Questions:

1. Should I switch from body (includes) to regex for more precision?

2. How can I prevent false positives like praise comments getting removed?

3. Is there a clean way to allow common fan phrases (“like,” “watch,” “great cast”) without weakening the filter?

Any advice or examples would be super helpful, or how to improve it, I'm new too all this html coding things. Thanks!