The Agent Skills Directory

[SAFE]: The skill's primary function is to search across multiple repositories using JetBrains' experimental jbcontext CLI and the official GitHub CLI (gh). These are legitimate developer tools used for organization-wide code analysis.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with internal and external tools (e.g., jbcontext repos, jbcontext search, gh repo view). This execution is well-defined, scoped to the user's repository access, and matches the described purpose of the skill.
[DATA_EXFILTRATION]: There are no signs of data being sent to untrusted third-party domains. All data operations (searching and reading) occur within the organization's GitHub environment or through the authorized jbcontext service.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of repository contents and search results. While this constitutes an attack surface, the instructions guide the agent through a standard research workflow without granting unsafe capabilities like arbitrary file writing or persistence. No specific bypass patterns were identified.

org-search