platform-tracing-agentforce-configure
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs local file operations by reading an internal XML template (
assets/AgentforcePlatformTracing-template.xml) and generating a Salesforce metadata file. No network calls or external downloads are performed. - [DATA_EXFILTRATION]: No patterns for sensitive file access or data exfiltration were identified. The skill only interacts with standard Salesforce metadata project paths.
- [REMOTE_CODE_EXECUTION]: There is no evidence of remote script execution, package installation, or dynamic code execution.
- [PROMPT_INJECTION]: The instructions do not contain markers for overriding safety guidelines, bypassing constraints, or extracting system prompts.
- [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface where it takes a 'Desired state' from the user to populate the
{ENABLED}placeholder in the XML template. - Ingestion points: User input for enabling/disabling tracing in
SKILL.md. - Boundary markers: Absent.
- Capability inventory: Local file writing (
settings/AgentforcePlatformTracing.settings-meta.xml). - Sanitization: The instructions explicitly restrict the values to 'true' or 'false'. Given the limited capability of writing a singleton metadata file, the risk is minimal.
Audit Metadata