The Agent Skills Directory

[SAFE]: The skill defines instructions for a local development utility intended to audit Android builds for security and policy violations. No malicious command patterns, unauthorized network access, or persistence mechanisms were found.
[PROMPT_INJECTION]: A static detector flagged a potential concealment pattern. Analysis indicates this is a false positive triggered by technical descriptions of the tool's offline operation ('makes no API calls and needs no credentials') and instructions for the agent to report all findings to the user. The skill actually emphasizes transparency by requiring the agent to report 'hint' and 'ref' fields for every finding.
[COMMAND_EXECUTION]: The skill utilizes the gplay preflight CLI tool. This is consistent with the primary purpose of the skill to scan local build artifacts for hygiene and compliance. The instructions include a safety measure advising the agent to confirm flags via --help before execution.

gplay-preflight