The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the tessie-api CLI tool to perform vehicle management and telemetry tasks. This includes executing commands with physical impacts, such as honking the horn, flashing lights, and scheduling software updates, as specified in SKILL.md and manifest.yaml.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes external telemetry data, creating a potential surface for indirect injection from the vehicle's state information.
Ingestion points: Data is fetched via subcommands like status, state, and request as described in SKILL.md.
Boundary markers: The instructions do not define specific delimiters for separating vehicle telemetry data from the agent's prompt context.
Capability inventory: The agent can perform vehicle actions and write data to the API via tessie-api command and tessie-api request (documented in SKILL.md and manifest.yaml).
Sanitization: There are no explicit instructions for sanitizing or escaping the data received from the API.
[SAFE]: The skill's behavior aligns with its documented purpose of vehicle management. It utilizes official domains (api.tessie.com, tessie.com) for all operations and follows best practices for credential management by using a dedicated auth CLI flow and instructing the agent to never print sensitive tokens.

tessie