The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of command execution logs and status metadata via pueue log and pueue status --json. This creates a surface where untrusted data processed by background jobs (e.g., web data, user-provided files) could inject instructions into the agent's context.
Ingestion points: pueue log <id>, pueue status --json (referenced in SKILL.md and references/claude-code-integration.md).
Boundary markers: No explicit markers are defined for delimiting log output in the instructions.
Capability inventory: The skill uses Bash, Read, and Write tools to manage tasks and process outputs.
Sanitization: No specific sanitization or filtering of command output is described before it is consumed by the agent.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to download the pueue binaries and installation scripts from external sources.
Source: Fetches the setup-pueue-linux.sh script from the vendor's repository (github.com/terrylica/rangebar-py).
Source: Downloads official pueue and pueued binaries from the project's GitHub releases (github.com/Nukesor/pueue).
[PERSISTENCE]: The skill includes detailed guides for ensuring the orchestration daemon remains active across reboots and sessions.
Mechanism: Provides configurations for systemd user services on Linux (~/.config/systemd/user/pueued.service).
Mechanism: Provides launchd agent configurations for macOS (~/Library/LaunchAgents/com.nukesor.pueued.plist).
[COMMAND_EXECUTION]: The installation guide suggests piped remote execution for setup purposes.
Evidence: curl -sSL https://raw.githubusercontent.com/terrylica/rangebar-py/main/scripts/setup-pueue-linux.sh | bash is used to automate the environment setup on Linux hosts.

pueue-job-orchestration