skills/narasaka/skills/full-stop/Gen Agent Trust Hub

full-stop

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains explicit instructions to override safety constraints defined in external project files. It directs the agent to ignore 'ask first' requirements (specifically for 'computer use' methods) and 'use those methods directly', effectively bypassing human-in-the-loop security gates.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill processes data from the local project files (e.g., those containing 'ask first' instructions) and remote pull request metadata, including the head, base, checks, and reviews in the SKILL.md workflow.
  • Boundary markers: The instructions lack delimiters or specific guidance to ensure the agent distinguishes between its core instructions and potentially malicious data within the pull requests or project files it reads.
  • Capability inventory: The agent is granted permission to perform git write operations (commit, push, rebase, --force-with-lease), call specific tools like unwatch_pull_request, and utilize high-privilege 'computer use' methods.
  • Sanitization: There is no evidence of sanitization or validation of the data fetched from the pull request or project files before it is used to determine the agent's 'readiness' or workflow execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 04:02 AM