report-comment
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to interact with the local file system and Git repository.
- Evidence: Steps 1 and 3 utilize commands such as
mkdir -p,touch,git log, andgit diffto create report files and gather technical changes. The issue key derived from the environment is used to construct local file paths. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from external, untrusted sources including Git commit messages and Jira ticket descriptions. This creates a surface where an attacker could embed malicious instructions in code metadata or ticket content to influence the agent's behavior.
- Ingestion points: Untrusted data enters the context via
git log,git diff, and thegetJiraIssuetool from the Atlassian MCP server. - Boundary markers: The instructions do not define specific delimiters or "ignore" warnings to encapsulate the external data during processing.
- Capability inventory: The skill has the capability to write to the file system (
touch,mkdir) and execute repository queries (git). - Sanitization: No explicit sanitization or filtering of the external content is described before the data is analyzed by the model.
Audit Metadata