Novelty Causal Explanation Engine
Develops and stress-tests hard-to-vary causal theories for unexplained phenomena through multi-cycle evolutionary exploration.
- Repository
- danielmiessler/LifeOS
- Stars
- 19,682
- Licence
- MIT
- Files
- 5 files
What it does
Once installed, the agent stops giving quick guesses about mysteries and instead runs a structured process: it gathers observations, rival explanations and available tests, generates many strange candidate theories, mixes and mutates them over several rounds, and keeps only those that survive strict checks for being hard to vary and testable. Each surviving theory comes with its lineage, a risky prediction that could prove it wrong, and a note of what it still leaves unexplained. Speculative ideas are kept but clearly labeled as unproven.
Great for
- Researchers or thinkers working on an open question with no good accepted explanation yet
- People who want a ranked set of rival causal theories tested against known observations
- Teams running a time-boxed creative sprint to generate testable hypotheses
Skip it if
- Simple brainstorming or name/idea lists, where the single-pass BeCreative skill is the better fit (the skill itself says so)
- Questions with well-established answers, since the process is designed for phenomena not adequately explained today
- Medical, legal or safety decisions, because the output is speculative hypotheses that need expert validation and could be mistaken for established fact
Try asking
- “Nobody has a good account of why this species migrates at night. Develop a new theory.”
- “Here are three theories of the anomaly. Which one actually explains it?”
- “Give me 20 minutes of novelty on why the outage only hits on Tuesdays.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 5 original files, the MIT licence and an attribution note.
Novelty.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip Novelty.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip Novelty.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Provide the phenomenon, known observations, constraints, rival explanations and any tests you can actually run; thin inputs produce thin theories.
- Ask for the rapid or selection intent when you only need a quick filter or time-boxed session rather than a full multi-cycle run.
- Treat survivors as hypotheses to test, not conclusions, and check the stated residue and discriminating prediction yourself.
What's inside
- HardToVary.md
- Intents/EvolutionCycle.md
- Intents/ExplanatorySelection.md
- Intents/RapidCycle.md
- SKILL.md
Redistributed unmodified from danielmiessler/LifeOS under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.