The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted source material (user-provided articles or stories) as described in workflow.md (Step 1.2) to generate storyboards and image prompts. This creates an attack surface where instructions embedded in the source content could influence the agent's logic.\n
Ingestion points: Source content is saved to source.md and analyzed to produce analysis.md, storyboard.md, and subsequent image prompts.\n
Boundary markers: The skill does not implement specific delimiters or instructions to ignore embedded commands within the source material.\n
Capability inventory: The skill possesses the capability to write various files to the local file system, execute the merge-to-pdf.ts script, and invoke external image generation tools via npx.\n
Sanitization: No explicit sanitization or validation of the source text is performed before it is used in the narrative generation process.\n- [COMMAND_EXECUTION]: The skill relies on the execution of shell commands and local TypeScript scripts to perform its tasks.\n
Evidence: Instructions in SKILL.md and workflow.md direct the agent to run npx -y bun ${SKILL_DIR}/scripts/merge-to-pdf.ts and npx -y bun ${SKILL_DIR}/../baoyu-image-gen/scripts/main.ts.\n
System Tools: The skill also suggests using system utilities like pngquant, optipng, or sips for image compression in SKILL.md Step 7.1.

baoyu-comic