artifact-integrator
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection vulnerability surface detected within the artifact analysis workflow.
- Ingestion points: The skill ingests untrusted data from the
.claude/context/runtime/integration-queue.jsonlfile and reads the contents of arbitrary files listed in theaffectedFilesarray during backward-propagation validation (Step 3.5). - Boundary markers: The protocol does not specify the use of delimiters or 'ignore embedded instructions' warnings when handling content from external sources.
- Capability inventory: The skill has access to sensitive tools including
Bash,Write,Edit,TaskCreate, andTaskUpdate. - Sanitization: No sanitization, escaping, or validation logic is applied to the extracted
patternorrationalestrings before they are interpolated into task descriptions and integration reports.
Audit Metadata