ClawHub

Claude Prompt Engineering - 提示词工程大师

v1.0.0

系统性应用Claude源码的9条提示词工程规则，提升AI回答质量，确保准确、简洁、无编造，适配多场景任务需求。

⭐ 0· 32·0 current·0 all-time
byjjx@jx-76
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
The name/description claim to apply prompt-engineering rules and the SKILL.md contains only rules, templates, and a tool description for applying those rules. There are no unrelated binaries, credentials, or external services requested — consistent with a prompt-authoring skill.
ℹ
Instruction Scope
The instructions explicitly describe injecting a system prompt and applying rule sets to the current conversation (apply_rules). That's within the skill's purpose. Minor note: the SKILL.md also says it will 'check AUTO_APPLY_PROMPT_RULES at session start' and shows a config path (.openclaw/config.json), which means the skill may look at or suggest changing local agent config — this is scope-expanding but coherent with a prompt-management skill.
✓
Install Mechanism
No install spec and no code files — instruction-only. This is the lowest-risk install posture (nothing is downloaded or written by the skill itself).
ℹ
Credentials
requires.env is empty, but the SKILL.md references an environment variable (AUTO_APPLY_PROMPT_RULES) and a local config file. This is a minor inconsistency: the skill may read that env/config if present, but it does not require any secrets or external credentials.
✓
Persistence & Privilege
always:false and no elevated privileges requested. The skill's main effect is to inject system-level prompt text when applied; that can influence agent behavior for the session but it does not request permanent system-wide privileges or other skills' credentials.
Assessment
This skill is instruction-only and appears to do what it says: inject and apply prompt-engineering rules. Before enabling: (1) review the exact system-prompt text (it will be injected into session/system context) so you know what constraints will be enforced; (2) note the SKILL.md may check AUTO_APPLY_PROMPT_RULES and suggest edits to .openclaw/config.json — disable auto-apply/auto-inject if you prefer manual control; (3) there are no network calls or credential requests, but because system-prompt injection changes agent behavior globally for the session, consider testing in a non-critical conversation first.

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

latestvk975zrbzyss5trsm0vyrym4xe98431aa

License

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

Comments