The Agent Skills Directory

[SAFE]: No security issues or malicious indicators were identified. The skill acts as an orchestrator for an autonomous engineering workflow using internal tools.
[COMMAND_EXECUTION]: The skill sequences several engineering tools via slash commands (e.g., /ce:work, /ce:plan, /compound-engineering:test-browser) to automate code changes and testing. These tools are consistent with the skill's stated purpose.
[PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by ingesting user-supplied feature descriptions to drive autonomous implementation.
Ingestion points: User input via $ARGUMENTS in SKILL.md.
Boundary markers: Absent; user input is passed directly to the planning tool.
Capability inventory: High-privilege tools including /ce:work (file system modification) and /compound-engineering:test-browser (testing/execution) are present in the workflow.
Sanitization: Absent; the skill relies on the underlying agent's reasoning or tool-level validation to handle untrusted input.

lfg