The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to read and process external specification files which could potentially contain malicious instructions designed to influence code generation.
Ingestion points: The agent is instructed to read content from openspec/changes/<change-name>/, specifically proposal.md, tasks.md, and files within the specs/ directory.
Boundary markers: There are no explicit instructions or delimiters defined to isolate external data or to warn the agent to ignore natural language commands found within those files.
Capability inventory: The agent is authorized to perform file system operations, creating and modifying Flutter source code files in the feature/, data/, domain/, application/, and presentation/ directories, as well as updating routing configuration in app/router/.
Sanitization: The workflow does not include a step for sanitizing or validating the contents of the specification files before they are processed by the agent.
[COMMAND_EXECUTION]: The skill documentation references a shell command (npx skills add flutter/agent-plugins --skill '*') as a prerequisite. While this is presented as a text-based instruction for manual setup rather than an automated hidden execution, it involves the execution of external packages through a package runner.

flutter-coding