The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted screenplay data via the renku screenplay analyze context command. This content is interpolated into the agent's context without explicit boundary markers or instructions to treat it as data only, potentially allowing embedded instructions in the script to affect agent actions.\n
Ingestion points: SKILL.md and references/screenplay-analysis-cli-workflow.md ingest data via renku screenplay analyze context.\n
Boundary markers: Absent. The instructions do not provide clear delimiters or negative constraints to isolate screenplay text from agent instructions.\n
Capability inventory: The agent has the capability to execute renku CLI commands to read and write project files, as outlined in SKILL.md.\n
Sanitization: The skill mitigates risks by requiring structural validation through renku screenplay analyze validate before any analysis is persisted via renku screenplay analyze write.

screenplay-analyst