The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests untrusted content from GitHub issues and local documentation files (PRPs and PRDs).
Ingestion points: Content is read from docs/prps/, docs/prds/, and the output of gh issue view.
Boundary markers: The instructions do not define delimiters or specific warnings to prevent the agent from following instructions embedded within the processed data.
Capability inventory: The agent has access to Bash (executing CLI tools), Write (modifying local files), and Read tools.
Sanitization: There is no evidence of sanitization or filtering for the data extracted from issues or documentation before it is interpolated into new work-orders or used in further command execution.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform several operations using local command-line interfaces.
Executes gh CLI commands to interact with GitHub issues, including viewing, creating, commenting on, and labeling issues.
Executes git status and git log to analyze the current state of the repository.
Uses jq to dynamically modify local JSON tracking files (feature-tracker.json and manifest.json).

blueprint-work-order