The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data sources including repository files, command outputs, and CI logs. This creates an attack surface where malicious instructions embedded in these external sources could attempt to influence the agent's behavior during terminal execution.
Ingestion points: Git state, repository files, command output, and CI/build logs (SKILL.md).
Boundary markers: The instructions do not specify any delimiters or safety prompts to isolate external data from the agent's primary instructions.
Capability inventory: The skill has high-privilege access, including shell command execution, file modification, and git operations like committing and pushing (SKILL.md).
Sanitization: There are no explicit mentions of sanitizing, escaping, or validating the content retrieved from external sources before it is processed.
[COMMAND_EXECUTION]: The primary purpose of this skill is to execute terminal commands and perform git operations. Although these capabilities are required for the skill's intended use and are governed by guardrails (such as 'inspect before editing' and 'do not use destructive git commands'), they provide the agent with significant control over the local environment which must be monitored for misuse.

terminal-ops