rink-explainer-illustrations
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content such as articles, documents, and screenshots to generate illustration plans and prompts. This creates a surface for indirect prompt injection where instructions embedded in the processed data could attempt to influence the agent's behavior.
- Ingestion points: The skill reads untrusted data from articles, documents, and screenshots as specified in the 'Digest The Source' section of
SKILL.md. - Boundary markers: There are no explicit instructions to ignore or delimit embedded instructions within the ingested source material.
- Capability inventory: The agent has the capability to call an
image-generationtool and perform file-system write operations to save generated assets. - Sanitization: No explicit sanitization or filtering of the source content is mentioned before it is used to formulate prompts or file names.
- [DATA_EXFILTRATION]: The skill instructions in
SKILL.mdspecify hardcoded local filesystem paths for saving generated illustrations, including a path tied to a specific user profile:/Users/ayoungh/Developer/Character/rink/generated/explainers/<source-slug>/. While intended for workspace organization, this reveals internal directory structures and hardcodes a specific user's environment into the skill's logic.
Audit Metadata