The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources (Jira work items and Confluence pages) which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Commands such as acli jira workitem view, acli jira workitem search, acli confluence page view, and acli confluence blog view fetch content (summaries, descriptions, comments, page bodies) from Atlassian Cloud.
Boundary markers: The skill does not explicitly instruct the agent to use delimiters or specific boundary markers when processing retrieved content to distinguish it from system instructions.
Capability inventory: The skill allows execution of shell commands via the Bash tool and file system access (Read, Grep, Glob), which could be misused if an indirect injection is successful.
Sanitization: There are no explicit instructions for the agent to sanitize or validate the content retrieved from Jira or Confluence before acting upon it.

acli