The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the task files it processes.
Ingestion points: issues/prd-issue-{N}.json and issues/progress-issue-{N}.txt.
Boundary markers: Absent; the agent is not instructed to treat the content of these files as data distinct from instructions.
Capability inventory: The skill has extensive shell execution capabilities including git, npm, cargo, and python tools.
Sanitization: No sanitization is performed on the user stories or acceptance criteria before the agent uses them to guide its actions.
[COMMAND_EXECUTION]: The agent is instructed to execute a wide range of shell commands for testing, linting, and building projects. While standard for a coding assistant, these tools provide a significant attack surface if the agent is manipulated by untrusted content within the issue tracking files.

execute-tasks