The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, which creates a surface for indirect prompt injection.
Ingestion points: The skill uses the renku screenplay import-fdx command to import Final Draft files and takes user-supplied story briefs to generate screenplays, as documented in SKILL.md and references/screenplay-json-workflow.md.
Boundary markers: The instructions lack explicit boundary markers or directives for the agent to treat imported screenplay text as passive data, increasing the risk of the agent following instructions embedded within those files.
Capability inventory: The skill has the capability to write JSON files to the tmp/ directory and execute various shell commands via the renku CLI, such as renku create and renku screenplay apply.
Sanitization: There is no evidence of content sanitization or validation logic to filter out potential malicious instructions embedded in the imported FDX files before they are processed into screenplay JSON.
[COMMAND_EXECUTION]: The skill's primary functionality is driven by executing shell commands through the renku CLI.
Evidence: Multiple files, including SKILL.md and references/screenplay-json-workflow.md, list required commands such as renku project open, renku screenplay status, renku screenplay create, and renku screenplay apply to manage the project lifecycle and screenplay mutations.

screenplay-drafter