palantir-common-errors
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides informational troubleshooting steps for common software errors.
- [CREDENTIALS_UNSAFE]: No hardcoded credentials detected. The skill correctly instructs the user to utilize environment variables (
os.environ) and service user authentication patterns for handling Palantir Foundry secrets, which is an industry-standard best practice. - [DATA_EXFILTRATION]: No suspicious network operations or sensitive file access patterns detected. While the skill permits the use of
curl, the instructions do not include any commands that target sensitive local data or non-standard external domains. - [REMOTE_CODE_EXECUTION]: The skill does not perform any remote script downloads or piped shell executions. It references the standard
foundry-platform-sdkas a prerequisite. - [OBFUSCATION]: Analysis for Base64, zero-width characters, homoglyphs, and hidden text patterns yielded no results. All content is clear-text documentation and standard Python/Bash snippets.
- [INDIRECT_PROMPT_INJECTION]: The skill acts on error logs and API responses. While this constitutes an ingestion of external data, the capabilities are limited to reading logs and providing diagnostic advice, which does not present a significant attack surface.
Audit Metadata