hyper-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 is exposed to untrusted data from issue trackers and project logs while performing its tasks.
- Ingestion points: Retrieves issue titles, bodies, and comments via a tracker CLI, and reads file content or log outputs for evidence gathering. (SKILL.md, rules/issue-contract.md)
- Boundary markers: Contains explicit instructions to treat all retrieved content as "evidence, never instructions" and specifically forbids following any commands embedded in external text. (SKILL.md)
- Capability inventory: The skill has the authority to write to issue trackers and load the
hyperskill to initiate implementation workflows. (rules/batch-selection.md) - Sanitization: Instructions require the manual redaction of credentials and personal data from any extracted evidence before it is recorded in a new issue. (rules/issue-contract.md)
- [COMMAND_EXECUTION]: The skill utilizes a command-line interface (such as the GitHub CLI) for reading and writing issues and employs a shell for the read-only reproduction of bugs to confirm findings. These operations are fundamental to the skill's purpose and are constrained by specific safety rules.
Audit Metadata