executing-plans
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and execute instructions from external plan files, which establishes a surface for indirect prompt injection.
- Ingestion points: The agent is instructed to read a plan file in Step 1 of the SKILL.md file.
- Boundary markers: The skill does not specify the use of delimiters, XML tags, or other boundary markers to isolate potentially untrusted plan content from the agent's core instructions.
- Capability inventory: The skill facilitates the execution of arbitrary tasks described in the plan, which generally involves file system modifications and shell command execution using other available agent tools.
- Sanitization: There are no instructions provided to sanitize, filter, or validate the content of the plan file before execution; furthermore, the requirement to 'Follow each step exactly' increases the risk that the agent will obey malicious instructions embedded in a compromised plan file.
Audit Metadata