search-company-knowledge
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources such as Confluence pages and Jira issues, which constitutes an indirect prompt injection surface. Because these sources are typically controlled by internal employees, the risk is inherent to the search functionality rather than a specific vulnerability.
- Ingestion points: Untrusted data enters the agent context via the
getConfluenceContentandgetJiraIssuetools described in SKILL.md. - Boundary markers: The skill provides structural guidelines for synthesis but does not use explicit delimiters or 'ignore instructions' markers.
- Capability inventory: The skill leverages retrieval tools (search, getConfluenceContent, getJiraIssue); no high-risk write or network exfiltration tools are documented.
- Sanitization: The skill relies on the agent's synthesis process to filter and present information, with no explicit content sanitization mentioned.
Audit Metadata