The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands for Git operations (e.g., git worktree, git pull), GitHub CLI (gh), and interactions with Bitbucket and Jira APIs via curl. It also runs user-defined strings from the skills-config.yaml file for health probes (probe.command) and quality gates (qualityGateCommand). These executions are necessary for the skill's primary function of development orchestration.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from a project completion roadmap (markdown file) to determine work items. To mitigate the risk of the agent following instructions embedded in the roadmap data, it uses a deterministic selector script (select-next.mjs) to parse the file and provides specific constraints to sub-agents during dispatch.
[DYNAMIC_EXECUTION]: The skill utilizes execFileSync and spawn with sh -c to run health checks and test suites defined in the repository's configuration. These features are intended for advanced CI/CD-like automation and are properly scoped to the project environment.
[SAFE]: Authentication handling for Bitbucket, Jira, and GitHub follows best practices by using environment variables (e.g., BITBUCKET_ACCESS_TOKEN, JIRA_API_TOKEN). The skill also implements built-in redaction logic to ensure secrets do not leak into logs or implementation reports.

develop-batch