The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute local shell scripts (./scripts/new-wave.sh) and use the GitHub CLI (gh) to manage branches and pull requests. These operations are performed within the agent's local environment and are restricted to standard development tooling.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and acts upon task data from an external source (VibeKanban).
Ingestion points: Untrusted task data is fetched from the VibeKanban system using the mcp__vibe_kanban__list_tasks and mcp__vibe_kanban__get_task tools.
Boundary markers: The skill instructions do not define boundary markers or explicit safety warnings to ensure the agent ignores instructions embedded within the task descriptions.
Capability inventory: The agent has access to powerful tools including Bash (shell command execution), Write (file system modification), and gh (interaction with remote repositories), which could be misdirected if malicious instructions are present in the ingested data.
Sanitization: No evidence of data validation, filtering, or escaping is present in the workflow for processing external task content.

wave-workflow