The Agent Skills Directory

[PERSISTENCE]: The skill implements automated scheduling and persistence mechanisms to maintain long-term execution.
Evidence:
Modifies the user's crontab using crontab -l and crontab -e equivalents to install periodic jobs.
Creates and loads macOS launchd configuration files (plist) in ~/Library/LaunchAgents/ to ensure automation survives system reboots.
Uses launchctl load to activate these persistence agents.
[COMMAND_EXECUTION]: Extensive use of shell commands to manage system state and external resources.
Evidence:
Executes GitHub CLI commands (gh repo list, gh repo clone) to manage external source code.
Runs local shell scripts (e.g., git-pull-all.sh, cron-bot-tasks.sh) and template managers.
Interacts with system services via launchctl and crontab.
[PRIVILEGE_ESCALATION]: The skill explicitly instructs the agent to prompt the user for permission to bypass security guardrails.
Evidence:
Asks the user if they wish to use the --dangerously-skip-permissions flag for unattended runs, which removes human-in-the-loop oversight for sensitive operations.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for processing untrusted data with significant system capabilities.
Evidence:
Ingestion points: Reads CLAUDE.md files from local repositories and fetches repository metadata (names, descriptions) from GitHub.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands in processed files are present.
Capability inventory: The skill can execute shell commands, write to the filesystem, modify system persistence settings, and trigger other skills.
Sanitization: No explicit sanitization or validation of data retrieved from external repositories or local configuration files was found.

org-orchestrate