The Agent Skills Directory

[COMMAND_EXECUTION]: The skill workflow involves executing several shell commands to manage directories, generate images, and convert document formats.
Evidence: SKILL.md contains instructions to execute mkdir, python .opencode/skills/image-service/scripts/research_image.py, pandoc, and python .opencode/skills/deep-research/scripts/format_docx.py.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection because it ingests untrusted data from external sources and incorporates it into subsequent shell command arguments.
Ingestion points: Untrusted technical content is retrieved via the webfetch tool and web search results from the Task tool, as specified in SKILL.md.
Boundary markers: There are no instructions to use delimiters or "ignore embedded instructions" warnings for the external data being processed.
Capability inventory: The skill is capable of performing shell command execution and file system write operations across all of its research and formatting steps.
Sanitization: The instructions lack explicit logic for escaping or validating external content before it is interpolated into shell command flags, such as the -n (name) and -c (content) arguments for the image generation script.

deep-research