The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform network connectivity checks using ping and nc, interact with device services via the hdc tool, and execute local build scripts (./build.sh).- [DYNAMIC_EXECUTION]: The agent is instructed to generate C++ source code for unit tests, which is then compiled and executed on a target device. This is the primary function of the skill but involves runtime execution of binary content generated by the AI.- [INDIRECT_PROMPT_INJECTION]: The 'Automated Flow' logic in SKILL.md and references/execution.md directs the agent to read test output and modify source code to fix errors. This feedback loop is an indirect prompt injection surface where malformed test results could influence the agent's logic. Ingestion points: Test execution output via hdc shell commands. Boundary markers: None identified to isolate external tool output from instructions. Capability inventory: Access to Bash, Write, and Edit tools. Sanitization: No validation or filtering of device output is mentioned before the agent processes it.

openharmony-ut