The Agent Skills Directory

[DATA_EXFILTRATION]: The skill enables querying of gateway request traces which contain raw request and response payloads. These payloads are explicitly documented in the reference files as containing sensitive data. The skill provides clear instructions and warnings regarding the handling of these fields as part of its monitoring and observability functions.
[PROMPT_INJECTION]: The skill establishes an indirect prompt injection surface by ingesting untrusted data from gateway traffic logs. An attacker could potentially send malicious instructions through the gateway to influence the agent's behavior when it later analyzes recent traffic or errors.
Ingestion points: Spans query API (detailed in references/monitoring.md) and SQL queries on the traces table (detailed in references/observability.md).
Boundary markers: The skill does not currently implement specific delimiters or 'ignore' instructions for the data ingested from traces.
Capability inventory: The skill possesses the capability to modify platform configurations via tfy apply, perform network operations via curl, and interact with management APIs via tfy-api.sh.
Sanitization: No content filtering or sanitization is performed on trace payloads before they are presented to or analyzed by the agent.
[COMMAND_EXECUTION]: The skill utilizes the tfy CLI and a bundled bash helper script (tfy-api.sh) to perform management operations and query platform status. This command execution is appropriate for the skill's administrative purpose and is performed within the constraints of the execution environment.

truefoundry-gateway