Prompt Engineering Mastery
v1.0.0Comprehensive system for designing, testing, optimizing, and managing clear, role-aware, actionable, focused, and testable prompts for AI models.
⭐ 0· 319·0 current·0 all-time
by@1kalin
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description match the provided instructions: this is an instruction-only prompt-engineering system (CRAFT, 4-layer architecture, CoT, chaining). It requests no binaries, no env vars, and has no install spec, which is proportionate. Minor inconsistency: README's install name ('afrexai-prompt-engineering') and links reference AfrexAI context packs and external paid content, while the registry slug is 'afrexai-prompt-mastery' and source/homepage are unknown — a packaging/branding mismatch that looks sloppy and deserves verification.
Instruction Scope
The SKILL.md content is large and contains many templates and techniques appropriate for prompt engineering. However, the pre-scan detected prompt-injection patterns such as 'ignore-previous-instructions' inside SKILL.md. That pattern is not needed to teach prompt engineering and could be used to instruct an agent to bypass higher-priority system/safety instructions. The file also suggests reading attached files and retrieved context (normal), but there is no explicit, justified need to override system-level constraints — review SKILL.md for any phrases that tell the model to ignore system messages, safety guards, or prior instructions and remove them before use.
Install Mechanism
No install spec and no code files — the skill is instruction-only. This minimizes install-time risk (nothing downloaded or written to disk).
Credentials
The skill declares no required environment variables, no primary credential, and no config paths. That is proportionate for a documentation-style prompt-engineering skill.
Persistence & Privilege
Flags show always:false and user-invocable:true (defaults). The skill does not request persistent presence or system privilege. Because it can be invoked autonomously by the agent platform, the prompt-injection concern in the SKILL.md increases risk if autonomous invocation is enabled — see user guidance.
Scan Findings in Context
[ignore-previous-instructions] unexpected: A prompt-engineering guide should teach good prompting practices but does not need to include directives that tell a model to ignore prior/system instructions or safety constraints. Presence of this pattern is a potential prompt-injection attempt and should be removed or justified.
What to consider before installing
This skill appears to be a legitimate prompt-engineering guide and is low-risk in that it requires no credentials or installs. However: 1) Inspect the full SKILL.md for any lines that explicitly tell the model to 'ignore previous instructions', 'disregard system messages', or otherwise override safety/system context — remove those lines before enabling. 2) Because the skill can be invoked autonomously by the agent platform, do NOT enable autonomous invocation for this skill until you remove any prompt-injection instructions or you run it with a human-in-the-loop. 3) Verify the author/source before following external paid links in the README; the registry slug and README install name mismatch is a packaging inconsistency. 4) If you want to be extra cautious, test in an isolated agent instance or sandbox, and retain manual approval for prompts that could access external data or invoke tools. If you want, provide the full SKILL.md and I can point to the exact lines that triggered the injection pattern so you can edit them out.Like a lobster shell, security has layers — review code before you run it.
aivk97fh7g4wyjfdkeqcc084s7551820vbeclaudevk97fh7g4wyjfdkeqcc084s7551820vbeengineeringvk97fh7g4wyjfdkeqcc084s7551820vbegptvk97fh7g4wyjfdkeqcc084s7551820vbelatestvk97fh7g4wyjfdkeqcc084s7551820vbellmvk97fh7g4wyjfdkeqcc084s7551820vbeoptimizationvk97fh7g4wyjfdkeqcc084s7551820vbepromptvk97fh7g4wyjfdkeqcc084s7551820vbe
License
MIT-0
Free to use, modify, and redistribute. No attribution required.