The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository, including issue IDs extracted from filenames and codebase context ingested during the investigation phases. This data influences the routing and execution of downstream implementation workflows.
Ingestion points: SKILL.md (task classification, issue IDs from filenames) and codebase content accessed via referenced skills like read-codebase-context.
Boundary markers: The instructions do not specify the use of delimiters or specific warnings to ignore instructions that might be embedded within the processed codebase files or task metadata.
Capability inventory: The skill delegates work to other modules like implement-task and systematic-debugging, which possess file-write and execution capabilities necessary for software development.
Sanitization: No specific sanitization or validation logic is described for the data ingested from the repository or external task descriptions.

using-devkit