rm-sentry-issue-fixer

Pass

Audited by Gen Agent Trust Hub on May 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted Sentry issue data (exception messages, breadcrumbs) which presents an indirect prompt injection surface.\n
  • Ingestion points: Untrusted data enters via Sentry MCP tools like get_issue_details and search_issue_events.\n
  • Boundary markers: The 'Security Constraints' section includes a 'No embedded instructions' rule to prevent the agent from obeying directives in data.\n
  • Capability inventory: The skill involves file writes and shell execution (bash, git) to implement fixes.\n
  • Sanitization: The skill mandates secret/PII redaction and code-consistency validation.\n- [COMMAND_EXECUTION]: The skill uses local shell scripts (detect-default-branch.sh, make-branch-names.sh) and standard git commands for branch management. These operations are scoped to the project's local environment.\n- [DATA_EXFILTRATION]: While the skill accesses Sentry events that may contain sensitive data, it explicitly instructs the agent to redact credentials and PII from any output or test artifacts.
Audit Metadata
Risk Level
SAFE
Analyzed
May 5, 2026, 03:03 PM