executive-summary
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external project artifacts (YAML files) which serve as untrusted input. This creates a surface for indirect prompt injection where malicious instructions embedded in the artifacts could influence the agent. 1. Ingestion points: business-requirements.yaml, technical-requirements.yaml, and timeline.yaml located within the provided base_directory. 2. Boundary markers: The instructions do not include specific delimiters or warnings for the agent to treat the content of these files as data and ignore any embedded instructions. 3. Capability inventory: The skill performs shell operations (mkdir -p) and file writes to the local filesystem. 4. Sanitization: No sanitization, validation, or escaping of the YAML content is specified before it is interpolated into the final executive summary.
- [COMMAND_EXECUTION]: The skill uses a shell command to ensure the presence of an output directory, which could be exploited if path parameters are manipulated. Evidence: mkdir -p {base_directory}/summaries in the SKILL.md file.
Audit Metadata