The Agent Skills Directory

[SAFE]: The skill functions as a productivity tool for marketing and copywriting, consistent with its stated purpose. It prioritizes grounding content in existing project data to prevent hallucinations or inaccuracies.\n- [COMMAND_EXECUTION]: No instances of arbitrary shell command execution or subprocess spawning were detected. The skill uses standard file system tools (Read, Write) for managing product briefs and drafts.\n- [DATA_EXFILTRATION]: While the skill accesses project context files (such as '.ulpi/launch/positioning.md' and '.ulpi/design/DESIGN.md'), it does so within the scope of generating relevant copy. There is no evidence of sensitive credential access (e.g., SSH keys, environment secrets) or transmission of data to unauthorized external endpoints.\n- [EXTERNAL_DOWNLOADS]: No remote code execution patterns, external script downloads, or unpinned package installations were found. The skill does not rely on unverifiable external dependencies.\n- [PROMPT_INJECTION]: The instructions do not contain safety bypass markers, role-play jailbreaks, or attempts to extract system prompts. The skill focuses on adhering to specific platform compliance rules, such as avoiding upvote solicitation on Product Hunt.

launch-copy