skills/mukul975/anthropic-cybersecurity-skills/configuring-identity-aware-proxy-with-google-iap/Gen Agent Trust Hub
configuring-identity-aware-proxy-with-google-iap
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/process.pyexecutes system commands viasubprocess.runto interact with the Google Cloud SDK (gcloud). This is a standard method for programmatic cloud auditing and uses list-based arguments to avoid shell injection vulnerabilities.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically Google Cloud API responses and audit logs, which constitutes a potential surface for indirect prompt injection.\n - Ingestion points: External data is retrieved in
scripts/process.pyandscripts/agent.pyviagcloudCLI and Cloud API calls.\n - Boundary markers: The skill does not employ specific delimiters or boundary markers to isolate ingested log data from surrounding instructions.\n
- Capability inventory: The skill has capabilities to execute CLI commands and write report files to the local file system.\n
- Sanitization: Data retrieved from audit logs (such as principal emails) is included in reports without specialized sanitization or filtering beyond standard JSON parsing.
Audit Metadata