triage-issue
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data (bug reports and error logs) which represents a theoretical surface for indirect prompt injection.
- Ingestion points: Processes user-provided bug descriptions and stack traces in the initial triage phase within
SKILL.md. - Boundary markers: No specific text boundary markers are defined for incoming text, but instructions direct the agent to place the raw logs inside Markdown code blocks within generated output.
- Capability inventory: Includes Jira capabilities to query and modify tickets (
searchJiraIssuesUsingJql,getJiraIssue,addOrEditJiraIssueComment,createJiraIssue). It does not use local shell access or unauthorized outbound network utilities. - Sanitization: Implements robust human-in-the-loop mitigation by strictly mandating user approval before executing any Jira modifications.
- [SAFE]: No signs of prompt injection, unauthorized data exfiltration, malicious persistence, credential harvesting, or obfuscated code were identified across the files.
Audit Metadata