The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (technical specifications, requirements, or PRDs) to generate task lists, creating a potential surface for instructions embedded in those documents to influence agent behavior.
Ingestion points: The skill analyzes content from specs, plans, or PRDs found in the environment or provided by the user (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions embedded within the source specifications.
Capability inventory: The skill utilizes file system write capabilities to create and update markdown files in the /tasks/ directory.
Sanitization: No sanitization or validation routines for the input text are specified.
[SAFE]: The inclusion of shell commands like git checkout and npx jest within the generated markdown output is consistent with the skill's purpose as a developer tool and does not represent an immediate execution risk by the skill itself.

task-generator