cursor-triage-issue-reports
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from Slack reports, replies, and technical artifacts which could be manipulated by an attacker to influence the triage outcome or tracker content.
- Ingestion points: Slack message text (
reporter wording), thread replies, log files, crash traces, and visual content from screenshots or videos. - Boundary markers: Includes a "Hard safety rules" section that enforces immutable source coordinates (
SOURCE_CHANNEL_ID,SOURCE_THREAD_TS) and mandates preflight checks before any write operation. - Capability inventory: The skill has capabilities to read Slack threads, read local files (logs/traces), read git repository history/changes, search and create issues in external trackers (e.g., Linear, GitHub), and post replies to Slack.
- Sanitization: Uses isolated, read-only workers for findings to ensure they cannot perform unauthorized Slack writes; however, the coordinator still processes their output into a final verdict.
Audit Metadata