The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) to list issues, labels, and project items, as well as to create new issues. These operations are essential for the skill's purpose and are performed using the agent's existing credentials.
[DYNAMIC_EXECUTION]: A Python one-liner (python3 -c ...) is used to parse JSON output from the GitHub CLI. This script is a static utility for data transformation and does not involve remote code execution or untrusted data evaluation.
[DATA_EXPOSURE]: The skill reads from the local CLAUDE.md file to determine routing logic. This access is localized to the project repository and does not involve exfiltrating sensitive system data.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from CLAUDE.md and user-provided task descriptions.
Ingestion points: Reads routing patterns from the ### Task Routing section of CLAUDE.md and processes task strings from user input.
Boundary markers: No explicit delimiters are used to isolate the configuration data from the skill's operational logic.
Capability inventory: The skill can search and create issues on GitHub using the gh tool.
Sanitization: No specific sanitization or validation of the keywords or repository names is performed, though the impact is limited to the skill's task-creation functionality.

task-routing