self-improving-question-system-for-better-results
v1.0.0Unlock dramatically better AI responses by automatically diagnosing and elevating prompt quality. Use this skill when the user wants more accurate, actionabl...
⭐ 0· 73·0 current·0 all-time
by@evafan7
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description match the instructions: the skill evaluates prompt quality and asks targeted clarification questions before producing a final answer. It does not request external services, binaries, or credentials that would be unrelated to this purpose.
Instruction Scope
The SKILL.md gives a clear 10-point checklist and a deterministic loop for clarification. Two items to note: (1) the 'Silent Evaluation Principle' instructs the agent to keep the analysis internal and not expose the evaluation process — this is a design choice (not a direct security issue) but reduces transparency to the user about what checks were applied; (2) the trigger rules include subjective language ('any request where prompt quality is clearly insufficient'), which could lead to broader invocation than strictly necessary. Both are functional/design concerns rather than incoherent or malicious behavior.
Install Mechanism
No install spec or code files — instruction-only. Nothing is written to disk and there are no downloads or third-party packages, which minimizes install-time risk.
Credentials
The skill requires no environment variables, credentials, or config paths. Requested access is proportional (none) to the stated purpose of asking clarification questions and responding.
Persistence & Privilege
always is false and there are no indications the skill modifies other skills or system settings. Autonomous invocation (disable-model-invocation=false) is the platform default and, by itself, is not a concern here.
Assessment
This skill is instruction-only and internally consistent with its goal: it will silently evaluate your prompt on a 10-point checklist and ask focused clarifying questions until the prompt meets its quality thresholds or you say 'PROCEED ANYWAY'. It asks for no credentials and installs nothing. Consider whether you are comfortable with the 'Silent Evaluation' behavior (the skill will not show its internal checks to you) and with the subjective trigger language (it may ask clarification more often than you expect). If you prefer full transparency, ask the agent to expose the checklist/results rather than keeping them internal before installing or enabling the skill.Like a lobster shell, security has layers — review code before you run it.
latestvk975ye80xzytnppxbh0b8bxwvd83jv7w
License
MIT-0
Free to use, modify, and redistribute. No attribution required.