The Agent Skills Directory

[SAFE]: The skill manages resources via the TrueFoundry API, using standard authentication mechanisms and following vendor-recommended practices.
[SAFE]: Sensitive information such as API keys is handled through environment variables or local .env files, with clear instructions and script logic to prevent accidental exposure in logs or shell history.
[SAFE]: The tfy-api.sh script includes security guardrails, including validation to prevent path traversal attacks (checking for .. in paths) and using array-based argument passing for shell commands to prevent injection.
[SAFE]: The skill uses the allowed-tools configuration to restrict the agent's capabilities, limiting shell execution to the specific provided helper script.
[SAFE]: Explicit safety instructions are included to warn the agent against ingesting prompt content from untrusted external sources without user review, mitigating potential indirect prompt injection risks.

truefoundry-prompts