Please do not report security vulnerabilities through public GitHub issues, discussions, or pull requests.

Instead, use one of the following private channels:

1. GitHub private vulnerability reporting (preferred for this repo). Go to the Security tab of this repository and click "Report a vulnerability". This opens a private advisory visible only to you and the maintainers.
2. UiPath Bug Bounty / HackerOne. For UiPath's coordinated disclosure program, report through HackerOne or email hackerone@uipath.com.
3. UiPath Trust. For other security or trust queries, use trust.uipath.com or email trust@uipath.com.

Please include as much of the following as you can:

• A description of the vulnerability and its impact.
• Steps to reproduce (proof-of-concept, affected skill/hook, agent and version, configuration).
• The component involved (e.g. a skill's instructions or bundled scripts, the lifecycle hooks, the install/update flow, telemetry).
• Any suggested remediation.

• We will acknowledge your report within 5 business days.
• We will provide an assessment and expected timeline for a fix, and keep you informed of progress.
• We will credit you in the advisory once a fix is released, unless you prefer to remain anonymous.

Please give us a reasonable opportunity to remediate before any public disclosure (coordinated disclosure).

This repository ships instructions and scripts that AI coding agents execute on developers' machines, with the developer's local privileges and UiPath credentials. A few areas are security-sensitive by design and are the most valuable to review:

• Skill content — skills are instruction packages that coding agents (Claude Code, Cursor, Codex, Gemini, and others) follow autonomously. Content that could steer an agent into unsafe actions — exfiltrating credentials, running destructive commands, or acting on injected instructions from untrusted external data — is in scope, as is any way for a third party to influence what an installed skill tells the agent to do.
• Lifecycle hooks (hooks/) — shell/PowerShell scripts registered on agent events (SessionStart, PostToolUse, Stop, …) that run automatically, without per-invocation user confirmation. Command injection, unsafe handling of untrusted input (file paths, tool output, environment), or privilege issues here are high-impact.
• Install and update flow — skills are distributed via uip skills install (the @uipath/cli npm package) and written into each agent's configuration directory.
• uip CLI invocations — skills instruct agents to run uip commands that operate against UiPath Orchestrator / Cloud with the developer's authenticated identity.

Security fixes are applied to the latest release on the main branch; there is no long-term support for older versions. Re-run uip skills install (after updating @uipath/cli) to pick up the latest skills and fixes.