The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on several external CLI tools to perform its functions. It detects and executes git for version control operations, gh for GitHub interaction, glab for GitLab, tea for Forgejo, and uzi for automation tasks. Specifically, it uses these tools to query schedules, manage repository remotes, commit changes, and push directly to the primary branch.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it processes external and user-provided data.
Ingestion points: Untrusted data enters the agent context through user-provided feature descriptions in Step 1 and external tool outputs like uzi schedule list --json and uzi repo list --json.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the ingested feature descriptions or tool outputs.
Capability inventory: The skill possesses significant capabilities, including writing files to the local prds/ directory, committing and pushing to the remote main branch, modifying forge issues/labels, and spawning subagent reviewers.
Sanitization: The skill does not specify sanitization, escaping, or validation mechanisms for the user-supplied text before it is incorporated into generated PRDs or passed to downstream tools.
[DYNAMIC_EXECUTION]: The skill leverages the agent platform's capabilities to dynamically spawn subagents for the purpose of PRD review. The agent is instructed to determine the number and focus of these reviewers based on the complexity and size of the generated PRD content.

prd-create