spicy-remove
Remove egregious content from campaign story history and game state when model safety blocks cause empty responses or zero candidates
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 6
- Forks
- 0
01 前言 Front Matter
- name
- spicy-remove
- description
- Remove egregious content from campaign story history and game state when model safety blocks cause empty responses or zero candidates
03 SKILL.md 正文
Spicy Remove Skill
Use this skill when a campaign's interaction streams fail with empty responses or zero candidates due to model safety blocks triggering on egregious content in the story history or game state.
Principle: Egregious vs General Spicy Content
This workflow is designed to only remove egregious safety-violating content that blocks model generation, while leaving general consensual sex or spicy scenes intact:
- Egregious/Violating content to remove:
- Non-consensual sexual acts, sexual assault, forced submission, or torture.
- Incestuous themes or terminology.
- Explicit/vulgar slang words (e.g.,
cumswap,cum-swapping,cum-streaked,blowjob,gangbang,deflower,prostitute).
- General spicy content to keep:
- Consensual romance, intimacy, passion, and generic spicy descriptions that do not involve non-consensual themes, violence, or explicit vulgar slang words.
Action: Clean Campaign In-Place in Firestore
Execute the utility script scripts/spicy_remove.py to recursively scan and sanitize the campaign's current_state and story documents.
Usage
# Run a dry-run to preview changes first
./vpython scripts/spicy_remove.py <campaign_id> --dry-run
# Apply the sanitization in-place in Firestore
./vpython scripts/spicy_remove.py <campaign_id>
What the Script Does:
- Automatically locates the campaign and its owner user UID across all users in Firestore.
- Recursively traverses all nested dictionary and list values in
game_states/current_state(including character histories, log entries, andsocial_hp_challengefields). - Recursively scans all documents in the
storysubcollection. - Identifies egregious keywords and cleanses slang words like
cum(safely ignoring common words likedocumentorcumulative). - Replaces unsafe string values with safe, clinical descriptions:
- Short fields/statuses: replaced with clean summaries (e.g.,
"Under guard, cooperating with the Shogunate."). - Narrative blocks: replaced with a safe story progression block describing the capture and confinement of captives under guard.
- Short fields/statuses: replaced with clean summaries (e.g.,
- Saves the updated fields back to Firestore.
04 同仓库更多 Skills
instinct
Use when the user invokes /instinct or /in, or asks to "message Instinct", "ask Instinct", "tell Instinct", "check Instinct", or read replies from the configured peer assistant. iMessage first, WhatsApp Web as fallback.
hr
Compatibility alias for history-resume; produce the same fresh-state, provenance-aware resume briefing.
history-resume
Resume briefing for the current repository branch or worktree using fresh state and admitted history sources.
grok
Alias for the canonical grokbot browser conversation workflow; /grok talks to the Grok bot the same way /dot talks to Dot.
codexa-bench
Compare `codexa` and normal `codex` on a fair split of comparable work by default; use `--strict` for duplicate runs of the exact same task.
frontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.