linear-walkthrough

Pass

Audited by Gen Agent Trust Hub on Jul 11, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill references an external technical blog for context, which is a common and safe practice for providing instructional background.
  • [PROMPT_INJECTION]: The skill processes untrusted external data such as source code, product documents, and meeting notes (Ingestion points: Intake section). It does not define specific boundary markers for this data (Boundary markers: absent) or include instructions for sanitizing or escaping the content (Sanitization: absent). While this represents a surface for indirect prompt injection, the risk is limited by the skill's focus on document generation using tools like file-reading and shell commands (Capability inventory: Intake section).
  • [COMMAND_EXECUTION]: The skill instructs the agent to use shell commands like cat and grep as secondary methods for reading source material if primary tools are not available. This is a standard capability for technical agents and is used here solely for information retrieval.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 11, 2026, 12:24 PM
Security Audit — agent-trust-hub — linear-walkthrough