refine-prompt
v2.56.0Transforms vague prompts into precise, structured AI instructions. Use when asked to refine, improve, or sharpen a prompt, do prompt engineering, write a sys...
⭐ 0· 134·0 current·0 all-time
byIlia Alshanetsky@iliaal
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (refine prompts) matches the SKILL.md: steps for assess → rewrite → validate and explicit rules for length, no invention, etc. There are no declared binaries, env vars, or installs that would be out of scope.
Instruction Scope
SKILL.md strictly limits actions to refining prompts and validating them; it does not instruct reading environment variables or contacting external endpoints. It does suggest offering to save results to a local file (.ai/PROMPT.md) but explicitly requires user confirmation before writing. This file-write offer is within the feature scope but is the only non-ephemeral side-effect to note.
Install Mechanism
No install spec and no code files — instruction-only. This is the lowest-risk install profile (nothing is written to disk by the installer because there is no installer).
Credentials
Requires no environment variables, credentials, or config paths. The single filesystem path mentioned (.ai/PROMPT.md) is reasonable for a prompt-saving feature and the skill explicitly asks for user confirmation before writing.
Persistence & Privilege
always:false and no declared persistence. The metadata allows model invocation (default), which is normal for skills; there is no indication the skill will persist or modify other skills. The only persistence behavior described is optional appending to .ai/PROMPT.md with explicit user approval — confirm before allowing writes.
Assessment
This skill appears coherent and low-risk: it refines prompts and asks for nothing it doesn't need. Before installing/using, be aware that it may offer to save refined prompts to a local file (.ai/PROMPT.md) — it states it will only write with your confirmation, so deny file writes if you prefer not to store prompts. As with any skill, watch the first few runs for unexpected behavior (unexpected file writes, requests for credentials, or outbound network calls). If you want extra caution, test it in a sandbox or deny the save-to-disk action until you trust the behavior.Like a lobster shell, security has layers — review code before you run it.
latestvk972237xgkp42w8d4ry3anpay584v1nj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.