skills/sebas-aikon-intelligence/antigravity-awesome-skills/cc-skill-project-guidelines-example/Gen Agent Trust Hub
cc-skill-project-guidelines-example
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The example
analyze_with_claudefunction inSKILL.mddemonstrates a vulnerability where untrusted content is passed directly into an LLM prompt. - Ingestion points: The
contentparameter in theanalyze_with_claudefunction (line 125) allows external data to enter the agent's context. - Boundary markers: The implementation lacks clear delimiters (such as XML tags or triple backticks) to isolate the untrusted
contentfrom the rest of the message instructions. - Capability inventory: The skill specifies capabilities for interacting with the Anthropic API, Supabase database, and Google Cloud Run environment.
- Sanitization: The provided code pattern does not include input validation or filtering to mitigate the risk of malicious instructions embedded within the processed content.
Audit Metadata