Security audit
Security checks for vulnerabilities and agentic risk
This skill is a Markdown-only report-formatting helper with minor quality issues but no privileged behavior, persistence, or hidden data access.
Install only if you want help standardizing final reports and review deliverables. Be aware that broad trigger words may cause it to load for general reporting tasks, and ignore or correct the undefined --help verification steps unless the publisher documents an actual command.
The trigger keywords and auto-load conditions are broad enough that the skill may activate in many unrelated contexts involving reporting or organization. In an agent setting, overly broad activation can override or interfere with more context-appropriate skills, increase accidental invocation frequency, and nudge workflows toward premature finalization or unnecessary evidence-handling steps.
The documentation for formatting findings includes a verification step to 'Run the command with --help flag,' but this file defines no command, executable, or operational interface. That instruction contradicts the actual content of the skill, which is a Markdown process/template for structuring outputs rather than a runnable command.
This section again instructs users to verify availability by running a command with --help, but no command is defined anywhere in the skill. The surrounding content describes manual report formatting and action-item generation, so the command-verification note actively misrepresents how the skill is used.
The appendix guidance says to verify the skill by running a command with --help, but the file contains only documentation and templates for producing structured output. This is an intent-code divergence because the documentation implies a runnable command interface that the skill does not actually provide.
No suspicious patterns detected.
