The Agent Skills Directory

[SAFE]: The skill primarily consists of instructional content and architectural guidelines for planning tasks. It includes a benign validation script (scripts/validate.sh) that uses grep to ensure the skill documentation adheres to a specific format.
[SAFE]: References to the AgentOps CLI (ao) and external engineering skill repositories are purely informational and do not involve unauthorized network requests, command execution, or remote downloads.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process external 'intent' and 'supplied text', which creates a surface for indirect prompt injection if the inputs contain malicious instructions.
Ingestion points: SKILL.md (Workflow Step 1) specifies reading accepted intent and supplied text from external sources or conversation history.
Boundary markers: The skill does not define explicit boundary markers or instructions for the agent to ignore commands embedded within the processed intent.
Capability inventory: The skill documentation mentions the use of AgentOps CLI tools (ao) and executes a validation script (scripts/validate.sh) for its own metadata.
Sanitization: There are no specific instructions for sanitizing or validating external input before it is shaped into an intent source.

plan