The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from several files within the repository which could be modified by a malicious contributor or in a pull request to influence agent behavior.
Ingestion points: Reads requirements.md, design.md, tasks.md, and docs/agents/verify.md (SKILL.md sections 1 and 'Cold-start control').
Boundary markers: Absent. The skill does not instruct the agent to distinguish between data content and potentially malicious instructions embedded in these files.
Capability inventory: The agent is empowered to execute shell commands (Launch/Doctor), drive system validation, and interact with web surfaces via kimi-webbridge.
Sanitization: Absent. The skill instructs the agent to 'Launch and Doctor from that file', implying direct execution of instructions found in docs/agents/verify.md without validation.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to launch and verify the application based on configurations found in docs/agents/verify.md or project.md. This autonomous command execution relies on the integrity of the files in the repository.

validate-feature