ClawHub

Prompt Optimizer

v1.0.0

When user asks to improve prompt, optimize prompt, better prompt, fix prompt, rewrite prompt, prompt engineering, make prompt better, enhance prompt, prompt...

⭐ 4· 448·0 current·1 all-time
byManish Pareek@mkpareek0315
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
✓
Purpose & Capability
Name and description match the behavior in SKILL.md: the skill optimizes prompts and offers templates, diagnosis, and a local prompt library. It does not request unrelated credentials or binaries.
ℹ
Instruction Scope
The runtime instructions include creating ~/.openclaw/prompt-optimizer and reading/writing settings.json, library.json, and history.json. Those file operations are consistent with saving prompts/history, but the SKILL.md also contains many 'You are ...' / system-prompt style templates (expected for a prompt tool). Review those templates because system-style content can influence agent behavior.
✓
Install Mechanism
No install spec and no code files — lowest-risk form (instruction-only). Nothing is downloaded or written beyond the explicit data directory the skill instructs to create.
✓
Credentials
The skill declares no environment variables or external credentials. The SKILL.md metadata asks for read/write tool permissions (to manage the local library), which is proportionate to its stated purpose.
ℹ
Persistence & Privilege
always:false and no autonomous elevation. The skill persists user data under ~/.openclaw/prompt-optimizer (settings, history, library). This is expected but does create persistent files in the user's home directory that you may want to inspect or manage.
Scan Findings in Context
[system-prompt-override] expected: The skill legitimately includes system-prompt style templates and 'You are ...' role instructions because it generates system prompts and templates. The scanner flagged this as a potential prompt-injection pattern — this is expected for a prompt optimizer, but you should still review any supplied templates for unexpected instructions or attempts to change agent-level behavior outside the skill.
Assessment
Before installing: (1) Review the SKILL.md (especially any system-prompt or template examples) to ensure there are no unexpected instructions that could alter agent behavior beyond prompt optimization. (2) Be aware the skill will create and persist files in ~/.openclaw/prompt-optimizer (settings.json, library.json, history.json); inspect their contents and delete them if you stop using the skill. (3) If your prompts contain sensitive data (passwords, API keys, proprietary text), avoid saving them in the skill's library; prefer ephemeral use. (4) Confirm that the 'read'/'write' permissions the platform grants the skill are limited to the intended directory; if possible, sandbox or restrict filesystem access to that path. (5) The pre-scan flagged system-prompt content — that is expected here, but if you see any templates that request network access, credentials, or to execute programs, do not use the skill and remove it.

Like a lobster shell, security has layers — review code before you run it.

latestvk979kgrmgbyf3bee5n899w7vbx81p230

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

⚡ Clawdis

Comments