litigation-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect prompt injection as it is designed to process untrusted external data such as court judgments, transcripts, and evidence lists. A malicious document could theoretically contain instructions meant to influence the agent's behavior during the analysis phase.
- Ingestion points: Document content is ingested via file paths (e.g., @path.md) or direct text pasting into the agent's context as described in SKILL.md.
- Boundary markers: The provided templates (e.g., template-internal.md, template-client.md) do not explicitly define strict data delimiters or instructions for the model to disregard embedded commands in the source text.
- Capability inventory: The skill utilizes file-writing capabilities to save analysis reports (as specified in references/hearing-transcript-extraction.md) but does not possess network or shell execution capabilities.
- Sanitization: There are no prescribed sanitization or escaping mechanisms for the input text before it is processed by the AI model.
Audit Metadata