implement
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted data from Linear issue trackers, creating a surface for indirect instructions.
- Ingestion points: The agent reads external Linear issue titles, descriptions, comments, and linked criteria (
SKILL.md,references/intake-and-handoff.md). - Boundary markers: The instructions do not specify formal delimiters for external data but rely on a structured 'review-correction loop' (
references/review-loop.md) and mandatory user validation for critical steps like opening Pull Requests. - Capability inventory: The skill has the capability to write to the file system (code and documentation), execute Git commands (commit, push, branch management), and update Linear issue statuses.
- Sanitization: The skill includes explicit directives to redact secrets, private details, and sensitive data before external publication (
references/delivery.md,references/intake-and-handoff.md).
Audit Metadata