The Agent Skills Directory

[SAFE]: The skill uses established communication channels (GitHub CLI, Bitbucket REST API, and Jira REST API) for issue management.\n- [SAFE]: It employs a sophisticated credential redaction mechanism in references/defer-mutation.js to prevent sensitive information from appearing in logs or deferred action records.\n- [SAFE]: Subprocess calls and shell commands are constructed using secure methods, such as temporary files for large text bodies and JSON builders (jq) for API payloads, mitigating prompt and command injection risks.\n- [SAFE]: The configuration parsing logic in references/read-config.sh is hardened against Python path poisoning attacks.

create-issue