eli5
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and process data from the user's repository, including manifests, configuration files, and source code, to generate explanations. This creates a surface where malicious instructions embedded in the codebase (e.g., in a README or code comment) could attempt to influence the agent's behavior.
- Ingestion points: Ingests repository content through recursive grep search, manifest analysis, and subagents (
Explore,general-purpose) reading specific files. - Boundary markers: The instructions do not explicitly mandate delimiters or specific instructions to ignore embedded prompts in the analyzed data, though they emphasize groundedness in real traces.
- Capability inventory: The agent can search the repository, read files, and launch subagents to process information.
- Sanitization: No explicit sanitization or filtering of external repository content is specified before the agent processes it for the explanation.
Audit Metadata