astro-observation-report
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external observational results and 'Uploaded PDF artifacts' to generate a structured manuscript. This workflow creates an inherent attack surface where malicious instructions could be embedded in the input data to influence the agent.
- Ingestion points: The skill processes user-provided event data (masses, spins, tidal parameters) and external PDF reference sources as defined in
SKILL.md. - Boundary markers: No specific boundary markers, delimiters, or 'ignore embedded instructions' warnings are provided to separate the instruction set from the untrusted data being analyzed.
- Capability inventory: While the provided skill files are Markdown only, an agent using this skill is expected to perform complex document generation tasks, including file system access and potentially executing plotting scripts (as implied by the figure list in
references/observational_results_patterns.md). - Sanitization: The skill contains no instructions for sanitizing or validating the content of the external PDFs or numerical data before incorporation into the context.
Audit Metadata