sarif-issue-reporter
Warn
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: MEDIUMINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes SARIF files provided by the user, which are untrusted external inputs. A malicious SARIF file could contain instructions within its fields designed to manipulate the agent's behavior during analysis or report generation.\n
- Ingestion points: The agent parses the SARIF file at the provided path in Phase 1 of the skill workflow.\n
- Boundary markers: The skill lacks explicit boundary markers or instructions to ignore embedded commands within the SARIF data.\n
- Capability inventory: The skill allows reading files, calculating security metrics, and generating reports.\n
- Sanitization: No input sanitization is performed on the text fields of the SARIF file.\n- [DATA_EXFILTRATION]: The skill and its helper script (scripts/sarif_helper.py) can read local files based on paths specified in the SARIF input. A malicious SARIF file could include paths to sensitive files such as credentials or configuration files.\n
- Evidence: The read_code_context function in the helper script opens and reads files using paths from the artifactLocation URI without validation.\n
- Context: While file reading is necessary for source code analysis, the absence of path restrictions could lead to the unauthorized exposure of sensitive information in the output report.
Audit Metadata