skills/udecode/plate/slate-ar-ship/Gen Agent Trust Hub

slate-ar-ship

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes external data, including user-supplied arguments and repository source code, which introduces a potential surface for indirect prompt injection. Maliciously crafted content within the files being reviewed could attempt to influence the agent's behavior during automated review or proofing steps.
  • Ingestion points: The skill reads from $ARGUMENTS and files in the "current tree," including temporary directories like .tmp/slate-v2.
  • Boundary markers: The instructions do not define explicit delimiters to separate untrusted file content from the agent's instructional context.
  • Capability inventory: The skill has the ability to execute git commands, run local scripts (e.g., finalize-autoresearch.mjs), and invoke companion skills (slate-ar-finalize, slate-ar-gate).
  • Sanitization: No specific filtering or sanitization of the ingested code is described, relying on the agent's internal guardrails and the requirement for human oversight.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands (git) and executes local JavaScript scripts (e.g., finalize-autoresearch.mjs) to manage the repository state. These actions are standard for developer tooling and are designed with safeguards that require plain-language user approval before executing permanent mutations like commits or stashing files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 05:10 PM
Security Audit — agent-trust-hub — slate-ar-ship