The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various development tools via the shell, including nx for task orchestration, testing, and linting, and prisma for database migrations. It also runs a local bash script, advance-pipeline-lock.sh, to manage the state of the development pipeline.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks because it reads and processes external data (story/task markdown files and existing project source code) which directly informs its development decisions and command execution.
Ingestion points: Consumes markdown documents (story.*.md, task.*.md) and project source files discovered by the Explore subagent.
Boundary markers: Lacks explicit boundary markers or safety instructions to distinguish between the agent's instructions and potentially malicious content within the files it reads.
Capability inventory: Possesses significant capabilities including shell command execution (nx, prisma), file system modification, and git branch management.
Sanitization: Does not perform sanitization or validation of content ingested from the project files before utilizing it in prompts or implementation logic.

develop