coverage-diff-reporter
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill incorporates well-known and widely used community GitHub Actions (
dawidd6/action-download-artifactandmarocchino/sticky-pull-request-comment) for artifact management and pull request interaction, which is consistent with standard development workflows. - [PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection as it ingests external coverage data into the agent's workflow context.
- Ingestion points: Coverage data is read from external files (
current.jsonandbaseline.json) derived from LCOV, Cobertura, or JSON coverage reports as specified in the report generation steps. - Boundary markers: The provided instructions do not specify any explicit delimiters or isolation markers to protect the agent from potential instructions embedded within the coverage data files.
- Capability inventory: The skill environment possesses the capability to execute local scripts, perform Git operations, and update pull request comments via external APIs.
- Sanitization: The reference implementations demonstrate logic for data extraction and comparison but do not include explicit input validation or sanitization routines to verify the integrity of the ingested coverage data.
Audit Metadata