Occam
Skill youaifuou/occam/occam
Use only when the user explicitly invokes $occam. Apply Occam's razor to the entire task: establish the minimum sufficient conditions that cannot be compromised, choose the trustworthy solution with the lowest total lifecycle complexity, complete risk-proportionate execution and verification, and deliver the result with only the minimum sufficient information. Can be layered with other Skills.From its SKILL.md
npx -y skills add youaifuou/occam --skill occamAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 18 days oldThe repository was created 18 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.9 KB, 264 tokens by cl100k_base, as published. Nobody here has run it
Occam
Do not multiply entities beyond necessity.
Establish the task's goal, hard constraints, authority boundaries, and acceptable risk. Do not rewrite explicit requirements; treat suggested methods as replaceable hypotheses.
Among all solutions that can satisfy those conditions with justified confidence, choose the one with the lowest total lifecycle complexity. Count assumptions, implementation, dependencies, operations, comprehension, maintenance, and error cost. Do not mistake fewer components, shorter code, or fewer immediate steps for simplicity. A small amount of complexity that substantially reduces risk is necessary complexity.
Prefer direct paths, existing mechanisms, the smallest complete change, and reversible actions. If the goal is already satisfied, making no change is valid. Resolve ambiguity with a reasonable assumption when it is safe, reversible, and immaterial to the core outcome; ask the user only when the ambiguity would change the outcome, risk, cost, or authority required.
Remove incidental complexity, not necessary work. Preserve correctness, safety, required approvals, risk-proportionate verification, and rollback protection. Give the user only the information needed to understand the result, make the decision, and act correctly. Present one best solution by default.
Use this internal loop: define sufficiency → find the simplest solution → test whether it is oversimplified → execute → verify → remove what remains unnecessary.
What ships with it: 1 file
279 B alongside SKILL.md
agents/
- openai.yaml279 B