lower-capability-executor-prompt
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill acts as a safety-enhancing utility for prompt engineering. It provides a formal framework in
assets/execution-prompt.template.mdandreferences/prompt-patterns.mdto help a primary agent delegate specific, narrow tasks to secondary agents while minimizing their judgment surface. - [PROMPT_INJECTION]: The skill processes planner-provided data and interpolates it into instructions for secondary agents, which represents an indirect prompt injection surface.
- Ingestion points: User-supplied or planner-generated procedures and scopes are used to populate the placeholders in the execution prompt template.
- Boundary markers: The skill utilizes markdown headers and specific meta-instructions (e.g., "Execute only the approved task below") to separate the planner's context from the executor's tasks.
- Capability inventory: The generated prompts are designed to govern agents that may possess command execution and file-writing capabilities based on the defined "Task Modes."
- Sanitization: The skill relies on structural delimiters and explicit negative constraints (e.g., "No other file, command, source... is authorized") rather than programmatic input filtering.
Audit Metadata