openspec-proposal-creation-cn

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands for discovery, directory creation, and structural validation (e.g., find, mkdir, grep, awk, bash). These commands include placeholders such as {change-id}, {capability-name}, and {proposed-id} which are intended to be replaced by user-supplied strings, presenting a potential path for command injection if input is not handled correctly by the agent.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes and validates specification files and proposals which could contain untrusted data from external sources.
  • Ingestion points: Files located in the spec/ directory are read and analyzed using grep, awk, and find commands, with the content then interpreted by the agent.
  • Boundary markers: The instructions do not specify any boundary markers or delimiters to separate the skill's operational instructions from the content of the processed specification files.
  • Capability inventory: The skill utilizes mkdir, grep, awk, ls, find, test, and bash to manage files and perform validation checks.
  • Sanitization: No explicit sanitization or filtering of the content within the specification or proposal files is performed prior to processing or execution of validation commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:05 PM
Security Audit — agent-trust-hub — openspec-proposal-creation-cn