Security audit
Security checks for vulnerabilities and agentic risk
This skill is mostly a prompt-design helper, but it asks for and advertises broad file, API, and shell-command capabilities without clear limits.
Review this before installing. It does not show malicious code or persistence, but it is under-scoped for the authority it requests. Use it only if you are comfortable granting file write and shell execution capability to a broadly triggered agent helper, and prefer restricting or disabling write/exec unless a task explicitly needs them.
The activation/usage description is broad and ambiguous, effectively suggesting the skill should be used whenever AI calls, orchestration, or LLM applications are involved. Overbroad triggers can cause the agent to apply this skill in contexts far beyond its narrowly described purpose, including contexts where write or exec-capable tooling is available.
The skill claims the free/LITE version is limited to prompt engineering and basic ReAct design, but later advertises file handling, API integration, and command execution. This inconsistency can mislead the agent or user about the skill's actual authority and operational scope, increasing the chance that powerful tools like write or exec are invoked unexpectedly.
The markdown advertises file writing, API integration, and command execution without prominent user-facing warnings or guardrails about side effects. In a skill environment where tools include write and exec, undocumented side-effectful capabilities materially increase the risk of unauthorized file changes, external calls, or shell actions being taken under the guise of a helper skill.
The document first says no extra API configuration is required and leaves API key configuration blank, then later says some features require API keys via environment variables. Contradictory setup guidance can cause unsafe secret-handling behavior, such as users pasting credentials into prompts or misconfiguring sensitive environment variables without clear boundaries.
The skill is presented primarily as a prompt/agent-design helper, yet it also claims command-execution support and includes troubleshooting for exec failures. That mismatch obscures the real risk profile of the skill, because users may enable or trust it without realizing it may drive shell execution through available tools.
