regression-suite
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and processes external design documents and bug reports which could contain malicious instructions designed to influence the agent's behavior during report generation.
- Ingestion points: design/gdd/systems-index.md, production/qa/bugs/.md, tests/**/, production/qa/smoke-*.md.
- Boundary markers: Absent. The skill does not define delimiters or specific instructions to ignore embedded commands in the source files.
- Capability inventory: Read, Glob, Grep, Write, Edit, AskUserQuestion, Bash.
- Sanitization: Absent. The skill does not mention escaping or sanitizing content extracted from external files before using it in the manifest or report.
- [DYNAMIC_CONTEXT_INJECTION]: The skill utilizes the dynamic context injection syntax to execute a local helper script when the skill is loaded.
- Pattern:
!bash "${CLAUDE_SKILL_DIR}/../../hooks/yaml-helper.sh" resolve_config --keys automation,workflow,qa.level,system_overrides - Evidence: This command is used to populate configuration variables used throughout the skill's logic.
- Context: The operation is a benign use of a local project-specific tool and is restricted by the allowed-tools policy defined in the skill's frontmatter.
Audit Metadata