Pipeline

Launch automated multi-skill pipeline that chains skills into a loop. Use when user says "run pipeline", "automate research to PRD", "full pipeline", "resear...

This skill appears to do what it claims (chain other skills and keep lightweight state/log files). Before installing or running it: 1) confirm you trust the downstream skills it will invoke (they may perform network access, write files, or use credentials); 2) be aware it will create/read/write .solo/pipelines/*.local.md and .solo/pipelines/*.log in your project or home directory (the metadata did not declare those config paths — expect file I/O); 3) it can run autonomously to chain stages, so always use the explicit confirmation step it requires and verify what it will do when asked; 4) to cancel a run, delete the state file named solo-pipeline-{project}.local.md; and 5) if you do not want terminal dashboards, use the recommended --no-dashboard flag or run only via the platform UI. If you want additional assurance, ask the author for a list of exact skill names this pipeline will invoke and inspect those skills for safety before use.

Type: OpenClaw Skill Name: solo-pipeline Version: 1.4.0 The skill bundle is suspicious due to two main vulnerabilities identified in `SKILL.md`. Firstly, it instructs the AI agent to execute external shell scripts (`solo-research.sh`, `solo-dev.sh`, `solo-pipeline-status.sh`) that are not provided within the bundle, introducing a significant supply chain risk as the content of these scripts is unknown and could be malicious. Secondly, the skill passes user-provided arguments directly to these external scripts via `Bash` without explicit sanitization instructions, creating a potential shell injection vulnerability (RCE risk) if the agent does not automatically quote or sanitize the input.