The Agent Skills Directory

[SAFE]: The skill performs read-only audits on UiPath projects using local CLI tools. It follows strict rules to never modify files and to report all findings with appropriate severity levels.
[PROMPT_INJECTION]: A static analysis alert regarding a 'DAN/jailbreak' attempt was determined to be a false positive. The referenced file references/agents/agents-lowcode-rules.md contains a rule designed to detect prompt injection in the audited project; it does not contain instructions for the AI agent itself to bypass safety filters.
[EXTERNAL_DOWNLOADS]: The skill fetches software development kit (SDK) documentation and validator catalogs from trusted vendor domains, including uipath.github.io and uipath.com. These operations are used exclusively for audit consistency and do not involve executing untrusted remote code.
[COMMAND_EXECUTION]: The skill uses shell commands to run official UiPath diagnostic utilities (e.g., uip rpa validate, uip agent review). The instructions explicitly prohibit the execution of business logic or side-effect-heavy commands like uip api-workflow run, ensuring the audit process remains safe and observational.

uipath-review