The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides shell command templates that the agent is instructed to execute. This includes running a binary bundled within the project repository at .opencode/bin/fluxctl and performing broad git operations such as git add -A and git commit --amend.- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and acting upon untrusted data from external files and project-specific task storage.
Ingestion points: The skill reads input from markdown specification files, user-defined idea text, and task state stored in the .flux/ directory (e.g., via fluxctl show and fluxctl cat).
Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between data and executable commands within these processed files.
Capability inventory: The skill enables shell execution for task management, invokes specialized subagents for memory and quality auditing, and triggers additional automated review skills.
Sanitization: The instructions do not specify any sanitization or validation of the data extracted from the external files before it is used in subsequent agent steps or shell command interpolation.

flux-opencode-work