The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the Bash tool to execute a local script tfy-api.sh and other commands like curl, uv, tfy, and pip for platform interaction and environment management.
[DATA_EXFILTRATION]: The skill manages highly sensitive credentials including TFY_API_KEY, Personal Access Tokens (PATs), and Secret Groups. Although the instructions mandate security policies such as token masking and one-time display, the capability to retrieve and transmit these credentials represents a data exposure surface.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests data from external API responses (e.g., workspace names, role descriptions) which are then interpolated into the agent context.
Ingestion points: API responses from GET requests to /api/svc/v1/workspaces, /api/svc/v1/roles, and /api/svc/v1/clusters are ingested as seen in SKILL.md and references/access-control.md.
Boundary markers: No explicit delimiters or boundary markers are defined in the instructions to separate external API content from agent instructions.
Capability inventory: The agent has extensive Bash tool access allowing for network operations (curl) and system commands.
Sanitization: There is no evidence of sanitization or validation of the data returned from the TrueFoundry API before it is processed by the agent.

truefoundry-platform