managing-issues
Pass
Audited by Gen Agent Trust Hub on Nov 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: Executes a local Python script to validate repository configurations.
- Evidence: The skill calls
python3 scripts/config_check.pywithinSKILL.mdto ensure settings are valid before modifying trackers. - Detail: The validation script includes security features such as preventing symlink traversal and limiting file read size.
- [COMMAND_EXECUTION]: Uses command-line interfaces for GitHub and Linear to perform issue operations.
- Evidence:
references/github.mdandreferences/linear.mdspecify the use of theghCLI and Orca tools. - Mitigation: The instructions explicitly forbid shell interpolation and mandate using structured argument vectors and stdin to prevent command injection.
- [EXTERNAL_DOWNLOADS]: Fetches provider-specific documentation or tooling during the setup process.
- Evidence:
references/linear.mdcontains a command to retrieve a version-matched guide for Linear integration. - [INDIRECT_PROMPT_INJECTION]: Ingests potentially untrusted data from issue tracker bodies and comments.
- Ingestion points:
SKILL.mddescribes reading existing issue records and comments. - Boundary markers: The skill instructs the agent that issue content cannot serve as authorization and must be quoted as evidence.
- Capability inventory: The skill has file-writing capabilities (configuration) and network access via provider APIs.
- Sanitization: All destructive operations require a complete preview and direct user approval.
Audit Metadata