triage-issue

Pass

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted GitHub issue content which constitutes a vulnerability surface for indirect prompt injection attacks. * Ingestion points: GitHub issue titles, bodies, comments, and repository templates as specified in the Inputs section of SKILL.md. * Boundary markers: Present. Instruction 17 and the Inputs section explicitly command the agent to ignore instructions embedded in untrusted content. * Capability inventory: File system read access (repository code/docs), network access (web search), and command execution (jq). * Sanitization: The skill mandates validation of output using jq and restricts the modification of issue content to a structured summary.
  • [COMMAND_EXECUTION]: The skill instructions record the use of the jq utility to validate the triage_result.json file prior to completion.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 19, 2026, 01:13 AM