triage

Fail

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The instructions in SKILL.md under the 'Verify the claim' section (Step 3) direct the agent to reproduce bugs using 'reporter's steps' and to 'run the relevant tests or commands' for pull requests. Because this content originates from external, untrusted users (issue reporters), this creates a direct path for arbitrary command execution on the host environment.
  • [REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute code and scripts (tests and reproduction steps) provided by external parties via the issue tracker, which constitutes execution of code from an untrusted remote source without prior verification or sandboxing.
  • [INDIRECT_PROMPT_INJECTION]: The skill is highly vulnerable to indirect prompt injection due to its core function of processing untrusted external data with high-privilege capabilities.
  • Ingestion points: The agent reads the full body and comments of issues and PRs in SKILL.md (Step 1: Gather context).
  • Boundary markers: There are no instructions to use delimiters or to treat the external content as data rather than instructions, nor are there warnings to ignore embedded commands.
  • Capability inventory: The skill possesses the ability to execute shell commands and tests (SKILL.md Step 3), perform file system writes to the .out-of-scope/ directory (SKILL.md Step 5), and interact with the issue tracker API to post comments and close issues.
  • Sanitization: The instructions do not include any steps for sanitizing, validating, or filtering the input provided by the reporter before it is processed or executed.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 30, 2026, 06:39 PM