The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted source material (text, file paths, or URLs) to generate comic storyboards and image prompts. This introduces a surface where malicious instructions embedded in the input could influence the agent's behavior during analysis or prompt construction.
Ingestion points: Source content is read and stored in source-{slug}.md and analysis.md during the setup phase (Step 1.1 in references/workflow.md).
Capability inventory: The agent is instructed to use the terminal for curl operations to download images (Step 7.2 in references/workflow.md) and uses write_file for outputting project files.
Boundary markers: The skill uses markdown headers and structured sections to organize content but does not implement explicit delimiters or 'ignore' instructions at all data interpolation points.
Sanitization: The skill includes a safety directive in SKILL.md to 'Strip secrets — scan source content for API keys, tokens, or credentials before writing any output file,' which mitigates the risk of credential exfiltration.

baoyu-comic