The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides installation instructions that involve downloading a shell script from the vendor's infrastructure (https://cli.runpod.net) and piping it to bash. This represents the official distribution method for the tool the skill is designed to manage.
[INDIRECT_PROMPT_INJECTION]: The skill instructions involve the agent ingesting and acting upon logs and outputs from external compute resources (pods and serverless workers) via runpodctl logs and runpodctl run. This creates a theoretical ingestion surface for data that could contain instructions, although the resources are typically managed by the user.
Ingestion points: Outputs from runpodctl pod logs, runpodctl serverless logs, and runpodctl serverless run documented in SKILL.md and reference/command-reference.md.
Boundary markers: Absent; the skill does not specify delimiters for data ingested from these logs.
Capability inventory: The skill is granted access to the Bash tool for the runpodctl command in SKILL.md, allowing infrastructure management (create, delete, update).
Sanitization: No sanitization is mentioned for the log content, though the skill advises using structured JSON parsing for command output.

runpodctl