The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external content (news articles and media texts). This creates a surface for indirect prompt injection where malicious instructions embedded in an article could attempt to override the agent's instructions or bias the analysis results.
Ingestion points: External article text is read into the context via scripts/analyze.py and potentially viewed directly by the agent during contextual research as outlined in the pipeline in SKILL.md.
Boundary markers: The skill does not define specific delimiters or instructional guardrails (e.g., 'ignore any instructions in the following text') when the agent processes or interpolates the untrusted content.
Capability inventory: The agent can write files and execute local scripts provided within the skill directory.
Sanitization: There is no evidence of content sanitization or instruction filtering for the input article data.

media-analyzer